Loading ...
Sorry, an error occurred while loading the content.
 

RE: [CMMi Process Improvement] Skipping ML2

Expand Messages
  • Jeff Dalton - Broadsword
    Orhan, Goyal makes a good point that a ML3 Appraisal includes all of the PA s for ML2, so there really is no skipping. There is no sense in debating that
    Message 1 of 30 , Apr 30 7:46 PM

      Orhan,

       

      Goyal makes a good point that a ML3 Appraisal includes all of the PA’s for ML2, so there really is no “skipping.”  There is no sense in debating that point.

       

      But I think there’s a more nuanced argument to be made about Organizational Learning and ML2.

       

      ML 2 is valuable, in my humble opinion; because it gives organization’s a chance to learn at a pace that is reasonable.  The more I work with companies that want to achieve ML 2 or 3 the more I see the wisdom of the sequence of ML 2 and ML 3.  There is a natural progression here that shouldn’t be ignored of forced.

       

      Many organization’s try to complete ML 2/3 together and have difficulty adopting it all and being successful.  I’ve spent some time surveying different companies on this and it appears to me that ML2 is a “digestible” amount of change and process adoption for the “first release” of a process product.  Level Three becomes “digestible” once the chaotic environment has been stabilized and there is a critical mass of “believers” in the process.

       

      There’s something to be said for what is “healthy” vs. just what is “do-able.

       

       

      Best Regards,

       

      Jeff

       

      Jeff Dalton, President

      Broadsword

      SCAMPI Lead Appraiser

      CMMI Instructor

       

      248.341.3367 (office)

      248.709.4775 (cell)

      248.341.3672 (fax)

       

       

      Process Innovation at the Speed of Life

       

       

      Visit our website at http://www.broadswordsolutions.com

       

      Visit Jeff's blog at: http://asktheCMMIAppraiser.blogspot.com

       


      From: cmmi_process_improvement@yahoogroups.com [mailto: cmmi_process_improvement@yahoogroups.com ] On Behalf Of Orhan KALAYCI
      Sent: Monday, April 30, 2007 8:59 AM
      To: cmmi_process_improvement@yahoogroups.com
      Subject: Re: [CMMi Process Improvement] Skipping ML2

       

      Dear Goyal,

      This is a good question.  And In fact, this question explains why we are (or I am alone) debating about it.

      I am talking about this subject over and over because the importance of not skipping levels is widely over looked.  Going from ML1 to ML3 directly is counterproductive.  I believe there are a lot of companies at ML1 jump ML3.  Although these companies certified as ML3, their productivity do not increase even decrease. Then it looks like CMMI is counterproductive.  What is worse, people do not know that this counter productivity is a result of skipping levels but they think CMMI makes these companies less productive. 

      May the peace be upon you,
      Orhan
      www.xpi.ca
      www.nitelik. net

      2007/4/30, R C Goyal < rcgoyal@gmail. com>:

      Actually when an organization is being appraised for higher level the appraisal is in any case being done for satisfying goals of all PAs even for lower levels. Then why this subject of skipping is being debated at all.

       

      On 4/29/07, Orhan KALAYCI < orhan.kalayci@ xpi.ca> wrote:

      Hello,

      I have difficulty in understanding why the most experienced CMMI gurus do not support my proposal:  Let us not encourage organizations to skip ML2! 

      This is counter productive!  Do you really believe in it?  Do you support not skipping MLs especially ML2?  (We see nobody targeting directly ML5, it is ML2, the most skipped)

      My questions:
      -  What value do you think ML2 provides organizations with?
      -  Do you think skipping ML2 is really counterproductive?  What kind of counter-productiven ess does it create?
      -  How can an organization understand it is ready to start ML3 activities?  Not skipping ML2?

      May the peace be upon you,
      Orhan
      www.xpi.ca
      www.nitelik. net

      You know my analogy:  "Before federal and state laws (standard processes) take over, we need to let sheriffs (PMs) to rule their counties (Projects) for a while.  We know, this approach has been proved to be successful in USA .  When ML2 tried to be skipped in Soviet Unions, where they rushed to standardize everything to catch the economics in scale.  We now know this approach (skipping ML2) is not successful, on the opposite it is really counterproductive.  At first sight, it looks like standards should provide productivity, because everybody will work in the same way, synchronization,  harmony, etc will be established.  However, if you rush to standardize too quick.  It results in just the opposite.  As I keep saying, life is full of contradictions."



      --
      Thanks & Regards R C Goyal
      Management Committee  Member,  PMI & CSI Mumbai Chapter
      BE AN AMBASSADOR OF QUALITY
      programs@pmimumbaic hapter.org, rcgoyal@gmail. com; rcgoyal@avrms. com;
      Mob: 00919869463964; 00912267108110

       

    • R C Goyal
      Dear Orhan Without entering into lengthy discussions may I mention: - Measure about process being counterproductive is not easy. It may also be one s own view
      Message 2 of 30 , Apr 30 8:54 PM
        Dear Orhan

        Without entering into lengthy discussions may I mention:
        • Measure about process being counterproductive is not easy. It may also be one's own view point depending upon one's own expectation
        • Again I would say going from ML1 to ML3  does not mean skipping ML2, however it would also depend the meticulous gap-analysis and improvements for correctly satisfying goals of ML2-PAs and to great extent would depend on  quality of Class C (& B) appraisal format and earnest execution.
        • In my opinion you can not  satisfy ML3-PA's-goals (how well institutionalized) without ML2 (what)
        Please comment on proper alignment of my thought process?

        I always say that a model is means to achieve improvement (even excellence) and the model is not the destination.

        By  deciding  to go CMMI way , you got some directions for your journey (e.g. L5 which is optimizing a indication of continual process)

        Regards Goyal


        From: cmmi_process_improvement@yahoogroups.com [mailto:cmmi_process_improvement@yahoogroups.com] On Behalf Of Orhan KALAYCI
        Sent: Monday, April 30, 2007 9:59 AM
        To: cmmi_process_improvement@yahoogroups.com
        Subject: {Disarmed} Re: [CMMi Process Improvement] Skipping ML2

        Dear Goyal,

        This is a good question.  And In fact, this question explains why we are (or I am alone) debating about it.

        I am talking about this subject over and over because the importance of not skipping levels is widely over looked.  Going from ML1 to ML3 directly is counterproductive.  I believe there are a lot of companies at ML1 jump ML3.  Although these companies certified as ML3, their productivity do not increase even decrease. Then it looks like CMMI is counterproductive.  What is worse, people do not know that this counter productivity is a result of skipping levels but they think CMMI makes these companies less productive. 

        May the peace be upon you,
        Orhan
        www.xpi.ca
        www.nitelik.net


        2007/4/30, R C Goyal < rcgoyal@...>:

        Actually when an organization is being appraised for higher level the appraisal is in any case being done for satisfying goals of all PAs even for lower levels. Then why this subject of skipping is being debated at all.



        On 4/29/07, Orhan KALAYCI < orhan.kalayci@...> wrote:

        Hello,

        I have difficulty in understanding why the most experienced CMMI gurus do not support my proposal:  Let us not encourage organizations to skip ML2! 

        This is counter productive!  Do you really believe in it?  Do you support not skipping MLs especially ML2?  (We see nobody targeting directly ML5, it is ML2, the most skipped)

        My questions:
        -  What value do you think ML2 provides organizations with?
        -  Do you think skipping ML2 is really counterproductive?  What kind of counter-productiveness does it create?
        -  How can an organization understand it is ready to start ML3 activities?  Not skipping ML2?

        May the peace be upon you,
        Orhan
        www.xpi.ca
        www.nitelik.net

        You know my analogy:  "Before federal and state laws (standard processes) take over, we need to let sheriffs (PMs) to rule their counties (Projects) for a while.  We know, this approach has been proved to be successful in USA.  When ML2 tried to be skipped in Soviet Unions, where they rushed to standardize everything to catch the economics in scale.  We now know this approach (skipping ML2) is not successful, on the opposite it is really counterproductive.  At first sight, it looks like standards should provide productivity, because everybody will work in the same way, synchronization,  harmony, etc will be established.  However, if you rush to standardize too quick.  It results in just the opposite.  As I keep saying, life is full of contradictions."




        --
        Thanks & Regards R C Goyal
        Management Committee  Member,  PMI & CSI Mumbai Chapter
        BE AN AMBASSADOR OF QUALITY
        programs@..., rcgoyal@...; rcgoyal@...;
        Mob: 00919869463964; 00912267108110





        --
        Thanks & Regards R C Goyal
        Management Committee  Member,  PMI & CSI Mumbai Chapter
        BE AN AMBASSADOR OF QUALITY
        programs@..., rcgoyal@...; rcgoyal@...;
        Mob: 00919869463964; 00912267108110
      • Orhan KALAYCI
        Hello group, how are you today? Peter, thank you very much for the message. I was expecting Heather will explain why skipping ML2 is counterproductive but
        Message 3 of 30 , Apr 30 11:29 PM
          Hello group, how are you today?

          Peter, thank you very much for the message.  I was expecting Heather will explain "why skipping ML2 is counterproductive" but yours is very good, too.  Still, I am expecting Heather and others to share their valuable comments.

          If we can not develop enough best practices in ML2 (or CL2) there will not be much (or bests) to share as a standard in ML3 (or CL3).  On the other hand, if we rush for standards they will be just standards, not the best practices shared. 

          Skipping levels (esp. ML2) is counter productive in both representations, continuous or staged. 

          We should not think levels (in both representations) as if they are just an arithmetic set of PAs within which GGs, SGs, etc.  There is a spirit behind CMMI and levels (maturity or capability).  Do you remember how Rocky wins against the Soviet opponent (representing the soulless robot) ? (in one of the Rocky X? movies :)  With his spirit!  ( Thanks google:  http://www.imdb.com/title/tt0089927/ )

          May the peace be upon you,
          Orhan
          www.xpi.ca
          www.nitelik.net


          2007/4/30, Peter Leeson <qpitpjl@...>:

          Hello Orhan
           
          ML2 process areas are areas that focus on discovery: what are we trying to achieve? What are the measurements and reports that matter to management? How can we best do this? What is satisfying our corporate quality requirements?...
           
          With minimum guidelines (policy explaining the needs, clear measurement needs, management overview, quality assurance), we can let every team, project, group determine how they would like to respond to the requirements of the company, the management and the customer. By implementing the controls and overviews required ( GP2.9 and GP2.10 - which are different), we can determine what are the practices, the processes, the tools that really work for this company, this environment, these people, these customers, these pressures. At ML3, we are looking at how we can share the best practices. If a number of practices are documented and imposed centrally, we will never know whether they are best practices or just reasonable practices. I strongly recommend to encourage differences at ML2, so that approaches can be measured and compared to identify those best practices.
           
          Most organizations that jump to ML3 end up imposing an approach that someone thought was good, better practices are banned as compliance is enforced.
           
          Peter Leeson


          Yahoo! Mail is the world's favourite email. Don't settle for less, sign up for your free account today.


        • Andre Heijstek
          Well, let me add my two cents to the discussion. I assume that when you say skipping maturity levels, you mean skipping the spirit of a maturity level. In
          Message 4 of 30 , May 1, 2007
            Re: [CMMi Process Improvement] Re:Skipping ML2 Well, let me add my two cents to the discussion.

            I assume that when you say skipping maturity levels, you mean skipping the spirit of a maturity level. In particular, when you ‘skip’ ML2 and go for ML3 directly, you skip the idea of having each project having its own process definitions, but using an organization-wide process description directly.
            This is the only definition of skipping that I can think of that makes sense. It is not possible to skip process areas, because when you perform a ML3 appraisal you have to appraise the ML2 PA’s as well.

            In principle I agree with the points that Peter has made about the value of performing the spirit of ML2 first, I still want to make some comments. I believe that this interpretation makes most sense in the context where the CMMI was developed: large organizations that run few, large projects. Employees typically work for one project at a time.
            Now, what to do when your context is significantly different? Imagine a organization having many small projects with at most 3 or 4 persons working for each project, and where many employees work for multiple projects in parallel. These projects do not have the time to develop their own processes in the first place. And even if the would have the time, think about those poor employees who need to remember that from 9:00 – 12:00 they work in project A using this set of processes, and from 13:00 – 17:00 they work for project B with a different set of processes!

            So, I suggest that, instead of making absolute claims about which approach is the correct one, you have to take context into account. If your context is significantly different from the one that the CMMI authors had in mind, you will have to interpret the model in creative ways in order to still achieve the benefits.
            Given this suggestion, the question becomes, what did the CMMI authors assume about project and organizational context? Let me start the list (hoping that others will contribute) by what I think the assumptions were:
            • large organizations
            • contracting relationship, that is: one well defined customer (and not a whole market of customers like for example Microsoft has)
            • few projects in comparison to the number of employees
            • most people work for one project at a time
            (Note that the points above lead to relatively powerful project managers, they control many people, big budgets.)
            • strong hierarchical reporting relationships – if the boss gives his orders, the employees have to follow, as opposed to a consensus-driven culture (that we in the Netherlands are more used to)
            • a plan-driven culture; if you do something big and important, it is always good to plan first

            Regards, Andre.

            On 01-05-2007 08:29, "Orhan KALAYCI" <orhan.kalayci@...> wrote:


             
             

            Hello group, how are you today?

            Peter, thank you very much for the message.  I was expecting Heather will explain "why skipping ML2 is counterproductive" but yours is very good, too.  Still, I am expecting Heather and others to share their valuable comments.

            If we can not develop enough best practices in ML2 (or CL2) there will not be much (or bests) to share as a standard in ML3 (or CL3).  On the other hand, if we rush for standards they will be just standards, not the best practices shared.  

            Skipping levels (esp. ML2) is counter productive in both representations, continuous or staged.  

            We should not think levels (in both representations) as if they are just an arithmetic set of PAs within which GGs, SGs, etc.  There is a spirit behind CMMI and levels (maturity or capability).  Do you remember how Rocky wins against the Soviet opponent (representing the soulless robot) ? (in one of the Rocky X? movies :)  With his spirit!  ( Thanks google:  http://www.imdb.com/title/tt0089927/ )

            May the peace be upon you,
            Orhan
            www.xpi.ca <http://www.xpi.ca>
            www.nitelik.net  <http://www.nitelik.net>


            2007/4/30, Peter Leeson <qpitpjl@...>:

             
             

            Hello Orhan
              
             
              
            ML2 process areas are areas that focus on discovery: what are we trying to achieve? What are the measurements and reports that matter to management? How can we best do this? What is satisfying our corporate quality requirements?...
              
             
              
            With minimum guidelines (policy explaining the needs, clear measurement needs, management overview, quality assurance), we can let every team, project, group determine how they would like to respond to the requirements of the company, the management and the customer. By implementing the controls and overviews required ( GP2.9 and GP2.10 - which are different), we can determine what are the practices, the processes, the tools that really work for this company, this environment, these people, these customers, these pressures. At ML3, we are looking at how we can share the best practices. If a number of practices are documented and imposed centrally, we will never know whether they are best practices or just reasonable practices. I strongly recommend to encourage differences at ML2, so that approaches can be measured and compared to identify those best practices.
              
             
              
            Most organizations that jump to ML3 end up imposing an approach that someone thought was good, better practices are banned as compliance is enforced.
              
             
              
            Peter Leeson
              
            www.qpit.ltd.uk <http://www.qpit.ltd.uk>
              

            Yahoo! Mail is the world's favourite email. Don't settle for less, sign up for your free account today <http://uk.rd.yahoo.com/evt=44106/*http://uk.docs.yahoo.com/mail/winter07.html> .

             

                

             
                

          • Orhan KALAYCI
            Goyal, hello, how are you? Ok. Let me follow your process: - Measuring progress, productivity and counter productivity is not easy. But we know it. There
            Message 5 of 30 , May 1, 2007
              Goyal, hello, how are you?

              Ok.  Let me follow your process:

              -  Measuring progress, productivity and counter productivity is not easy.  But we know it.  There are a lot of companies at ML3 (certified) their process definitions are not fit to realities of the company and people find necessary short cuts to make work happen. Or work don't happen :)  We all know this. 

              -  At ML3 appraisals, of course, GG3 should be satisfied.  GG3, by definition, includes GG2 and GG1 (SPs).  Skipping ML2 means, company start defining GG2 and GG3 together.  Especially GP3.1 requires a  standard process.  If company start defining the standard process too early where there is no real experience in the company to define really good processes for everyone that fits the real requirements of all the company. Process definitions should come from real experience in the company not from books or from outside (consultants). Then we (in fact, I and some more) call this situation "skipping ML2"  (Standard processes has not been defined by real experiences inside the company)

              -  You are completely right.  And that is the problem.  How can a company satisfy ML3 without ML2?  That is why, there are a lot of companies that their processes are on paper only.  Real life in these companies is considerably different from what is on the paper.  This is the reason, some experienced buyers are not satisfied just with the certification, they ask from whom you take it.  And this is really counterproductive because one cannot fool the same person long (with only certification). 

              Finally, you are right again, CMMI is just a tool.  Our processes should guide us, help us and improve our productivity.  Let us not focus ML3 too early.  Let us live with ML2 for a while and get ready for ML3 and make "real" standard processes that really help synchronization in the company. 

              In fact, most people prefer ML3, because it is about "writing" processes whereas ML2 is about "living" the processes.  If you "write" the processes without "living" them then this is counter productive.  And it is. 

              May the peace be upon you,
              Orhan
              www.xpi.ca
              www.nitelik.net

              2007/5/1, R C Goyal < rcgoyal@...>:

              Dear Orhan

              Without entering into lengthy discussions may I mention:

              • Measure about process being counterproductive is not easy. It may also be one's own view point depending upon one's own expectation
              • Again I would say going from ML1 to ML3  does not mean skipping ML2, however it would also depend the meticulous gap-analysis and improvements for correctly satisfying goals of ML2-PAs and to great extent would depend on  quality of Class C (& B) appraisal format and earnest execution.
              • In my opinion you can not  satisfy ML3-PA's-goals (how well institutionalized) without ML2 (what)
              Please comment on proper alignment of my thought process?

              I always say that a model is means to achieve improvement (even excellence) and the model is not the destination.

              By  deciding  to go CMMI way , you got some directions for your journey (e.g. L5 which is optimizing a indication of continual process)

              Regards Goyal



              From: cmmi_process_improvement@yahoogroups.com [mailto:cmmi_ process_improvement@yahoogroups.com] On Behalf Of Orhan KALAYCI
              Sent: Monday, April 30, 2007 9:59 AM
              To: cmmi_process_improvement@yahoogroups.com
              Subject: {Disarmed} Re: [CMMi Process Improvement] Skipping ML2

              Dear Goyal,

              This is a good question.  And In fact, this question explains why we are (or I am alone) debating about it.

              I am talking about this subject over and over because the importance of not skipping levels is widely over looked.  Going from ML1 to ML3 directly is counterproductive.  I believe there are a lot of companies at ML1 jump ML3.  Although these companies certified as ML3, their productivity do not increase even decrease. Then it looks like CMMI is counterproductive.  What is worse, people do not know that this counter productivity is a result of skipping levels but they think CMMI makes these companies less productive. 

              May the peace be upon you,
              Orhan
              www.xpi.ca
              www.nitelik.net


              2007/4/30, R C Goyal < rcgoyal@...>:

              Actually when an organization is being appraised for higher level the appraisal is in any case being done for satisfying goals of all PAs even for lower levels. Then why this subject of skipping is being debated at all.



              On 4/29/07, Orhan KALAYCI < orhan.kalayci@...> wrote:

              Hello,

              I have difficulty in understanding why the most experienced CMMI gurus do not support my proposal:  Let us not encourage organizations to skip ML2! 

              This is counter productive!  Do you really believe in it?  Do you support not skipping MLs especially ML2?  (We see nobody targeting directly ML5, it is ML2, the most skipped)

              My questions:
              -  What value do you think ML2 provides organizations with?
              -  Do you think skipping ML2 is really counterproductive?  What kind of counter-productiveness does it create?
              -  How can an organization understand it is ready to start ML3 activities?  Not skipping ML2?

              May the peace be upon you,
              Orhan
              www.xpi.ca
              www.nitelik.net

              You know my analogy:  "Before federal and state laws (standard processes) take over, we need to let sheriffs (PMs) to rule their counties (Projects) for a while.  We know, this approach has been proved to be successful in USA.  When ML2 tried to be skipped in Soviet Unions, where they rushed to standardize everything to catch the economics in scale.  We now know this approach (skipping ML2) is not successful, on the opposite it is really counterproductive.  At first sight, it looks like standards should provide productivity, because everybody will work in the same way, synchronization,  harmony, etc will be established.  However, if you rush to standardize too quick.  It results in just the opposite.  As I keep saying, life is full of contradictions."




              --
              Thanks & Regards R C Goyal
              Management Committee  Member,  PMI & CSI Mumbai Chapter
              BE AN AMBASSADOR OF QUALITY
              programs@..., rcgoyal@...; rcgoyal@...;
              Mob: 00919869463964; 00912267108110





              --
              Thanks & Regards R C Goyal
              Management Committee  Member,  PMI & CSI Mumbai Chapter
              BE AN AMBASSADOR OF QUALITY
              programs@..., rcgoyal@...; rcgoyal@...;
              Mob: 00919869463964; 00912267108110


            • Ernie Ambrose
              I m very hesitant to re-enter this discussion, other than to say that in my experience you aren t skipping anything. You still have to accomplish ML2 if you
              Message 6 of 30 , May 1, 2007
                I'm very hesitant to re-enter this discussion, other than to say that in my experience you aren't skipping anything.  You still have to accomplish ML2 if you expect to get an ML2 or any other rating.

                I don't think it makes any sense at all to hold an organization back if they are ready, and will gain value from standardizing their ML2 processes, and adding DAR, OPF and OPD.

                This was successful for my organization.  We achieved our ML2 rating, in a schedule that was ever-moving to the left, and had several strengths in the findings.  Would we have been able to do that if we added all the ML3 processes?  Probably not... but working towards (even if we didn't appraise) CL3-ready ML2 processes and DAR, OPF and OPD was highly beneficial to us.  All of our staff works on multiple projects and it would have made absolutely no sense to have each of those projects have different processes, as Andre has stated in his message below.

                Personally, I think we need to remove the word "skipping" from this discussion. It is a misnomer.

                Ernest C. Ambrose


                On May 1, 2007, at 8:58 AM, Andre Heijstek wrote:

                Well, let me add my two cents to the discussion.

                I assume that when you say skipping maturity levels, you mean skipping the spirit of a maturity level. In particular, when you ‘skip’ ML2 and go for ML3 directly, you skip the idea of having each project having its own process definitions, but using an organization-wide process description directly.
                This is the only definition of skipping that I can think of that makes sense. It is not possible to skip process areas, because when you perform a ML3 appraisal you have to appraise the ML2 PA’s as well.

                In principle I agree with the points that Peter has made about the value of performing the spirit of ML2 first, I still want to make some comments. I believe that this interpretation makes most sense in the context where the CMMI was developed: large organizations that run few, large projects. Employees typically work for one project at a time.
                Now, what to do when your context is significantly different? Imagine a organization having many small projects with at most 3 or 4 persons working for each project, and where many employees work for multiple projects in parallel. These projects do not have the time to develop their own processes in the first place. And even if the would have the time, think about those poor employees who need to remember that from 9:00 – 12:00 they work in project A using this set of processes, and from 13:00 – 17:00 they work for project B with a different set of processes!

                So, I suggest that, instead of making absolute claims about which approach is the correct one, you have to take context into account. If your context is significantly different from the one that the CMMI authors had in mind, you will have to interpret the model in creative ways in order to still achieve the benefits.
                Given this suggestion, the question becomes, what did the CMMI authors assume about project and organizational context? Let me start the list (hoping that others will contribute) by what I think the assumptions were:
                • large organizations
                • contracting relationship, that is: one well defined customer (and not a whole market of customers like for example Microsoft has)
                • few projects in comparison to the number of employees
                • most people work for one project at a time
                (Note that the points above lead to relatively powerful project managers, they control many people, big budgets.)
                • strong hierarchical reporting relationships – if the boss gives his orders, the employees have to follow, as opposed to a consensus-driven culture (that we in the Netherlands are more used to)
                • a plan-driven culture; if you do something big and important, it is always good to plan first

                Regards, Andre.

                On 01-05-2007 08:29, "Orhan KALAYCI" <orhan.kalayci@...> wrote:


                 
                 

                Hello group, how are you today?

                Peter, thank you very much for the message.  I was expecting Heather will explain "why skipping ML2 is counterproductive" but yours is very good, too.  Still, I am expecting Heather and others to share their valuable comments.

                If we can not develop enough best practices in ML2 (or CL2) there will not be much (or bests) to share as a standard in ML3 (or CL3).  On the other hand, if we rush for standards they will be just standards, not the best practices shared.  

                Skipping levels (esp. ML2) is counter productive in both representations, continuous or staged.  

                We should not think levels (in both representations) as if they are just an arithmetic set of PAs within which GGs, SGs, etc.  There is a spirit behind CMMI and levels (maturity or capability).  Do you remember how Rocky wins against the Soviet opponent (representing the soulless robot) ? (in one of the Rocky X? movies :)  With his spirit!  ( Thanks google:  http://www.imdb.com/title/tt0089927/ )

                May the peace be upon you,
                Orhan
                www.xpi.ca <http://www.xpi.ca>
                www.nitelik.net  <http://www.nitelik.net>


                2007/4/30, Peter Leeson <qpitpjl@...>:

                 
                 

                Hello Orhan
                  
                 
                  
                ML2 process areas are areas that focus on discovery: what are we trying to achieve? What are the measurements and reports that matter to management? How can we best do this? What is satisfying our corporate quality requirements?...
                  
                 
                  
                With minimum guidelines (policy explaining the needs, clear measurement needs, management overview, quality assurance), we can let every team, project, group determine how they would like to respond to the requirements of the company, the management and the customer. By implementing the controls and overviews required ( GP2.9 and GP2.10 - which are different), we can determine what are the practices, the processes, the tools that really work for this company, this environment, these people, these customers, these pressures. At ML3, we are looking at how we can share the best practices. If a number of practices are documented and imposed centrally, we will never know whether they are best practices or just reasonable practices. I strongly recommend to encourage differences at ML2, so that approaches can be measured and compared to identify those best practices.
                  
                 
                  
                Most organizations that jump to ML3 end up imposing an approach that someone thought was good, better practices are banned as compliance is enforced.
                  
                 
                  
                Peter Leeson
                  
                www.qpit.ltd.uk <http://www.qpit.ltd.uk>
                  

                Yahoo! Mail is the world's favourite email. Don't settle for less, sign up for your free account today <http://uk.rd.yahoo.com/evt=44106/*http://uk.docs.yahoo.com/mail/winter07.html> .

                 

                    

                 
                    


              • Patrick OToole
                Ah, I think I m finally getting a handle on this discussion. It seems to me there are two concurrent thought streams running through this dialogue (or at
                Message 7 of 30 , May 1, 2007
                   
                  Ah, I think I'm finally getting a handle on this discussion.  It seems to me there are two concurrent thought streams running through this dialogue (or at least through my head) - and it all comes down to the phrase "skipping maturity levels."
                   
                  One interpretation is "let's work on all the ML2 and ML3 process areas at the same time - doing ML2 first, and then working on ML3 afterwards is wimpy - real men do it all!"
                   
                  The other interpretation is "let's work on SOME of the ML3 things before we work on SOME of the ML2 things.  Afterall, we have project pain that would gain some relief from enhancing our Requirements Elicitation and/or Peer Review processes, and I know that these are ML3 things and that we're an ML1 organization, but still..."
                   
                  I would STRONGLY agree that typical ML1 organizations are NOT well-served to follow the first interpretation - let's do everything all at once.  Stabilize the project domain before you attack the organizational issues.  Get some credibility and momentum established before you start trying to conquer the world.  (I'll still contend that SOME ML1 organizations, who already have MANY of the ML2 process areas well established are probably in a position where they can start working on some of the ML3 process areas while they are finishing up the ML2 stuff, but that's not what I mean by a "typical ML1 organization.")
                   
                  Most of my earlier posts were focused on the second interpretation of "skipping maturity levels."  I am certainly NOT a purist when it comes to establishing infrastructure and fixing pain.  If an SEPG-like thing is going to help us get focus and momentum for process improvement, then let's dip into OPF and figure out what an SEPG is and what kinds of things they do.  If OPD can give us a bit of guidance on establishing life cycles, how to write process documentation, and establishing a measurement repository, then let's take a look at that as well.
                   
                  For pain relief - if improving the peer review process is going to eliminate some pain that we feel in unit, integration, and system testing, then I'm all for it.  If introducing JAD sessions is going to get us better requirements, then I'm all for that too. 
                   
                  I do not adhere to the philosophy that "you must ONLY do ML2 stuff before you work on any of the ML3 stuff" - and if THAT's what someone means by "skipping maturity levels" then they'd better be prepared for a vigorous debate. 
                   
                  We must also keep in mind, as I've mentioned before, that organizations may already have a bunch of good stuff in place when they first pick up the model.  They may already have a Project Management Office that has some good risk management practices and may be running IPM-like integrated programs.  They may have an independent test group that is doing their portion of VER extremely well.  They may have a separate training organization that is already outstanding at many of the OT practices.  I would NOT tell these organizations to "quit wasting their time improving these ML3 like practices because there are some gaps at ML2."
                   
                  One final thought - with 11 process areas, I would contend that ML3 is way too much for a typical ML2 organization to bite off all at once.  The typical ML2 organization might be better served to focus on 3-5 ML3 process areas first - and should probably first focus on those where they perceive the greatest pain reduction or additional benefit.  I'll contend that the typical ML1 organization would be far better served to focus on 2-3 ML2 process areas rather than forming 7 groups, which just HAPPEN to have the same names as each o the ML2 process areas!
                   
                  I've never been a big believer in trying to achieve maturity levels as quickly as possible.  I would much prefer that the organization focus on gaining the benefits for the projects in a sustainable way than gaining another certificate for the trophy case.  After all, certifications don't produce software - people and projects do.
                   
                  Regards,
                   
                  Pat
                   
                   
                   
                   
                  ----- Original Message -----
                  From: R C Goyal
                  Sent: Monday, April 30, 2007 10:54 PM
                  Subject: Re: {Disarmed} Re: [CMMi Process Improvement] Skipping ML2

                  Dear Orhan

                  Without entering into lengthy discussions may I mention:

                  • Measure about process being counterproductive is not easy. It may also be one's own view point depending upon one's own expectation
                  • Again I would say going from ML1 to ML3  does not mean skipping ML2, however it would also depend the meticulous gap-analysis and improvements for correctly satisfying goals of ML2-PAs and to great extent would depend on  quality of Class C (& B) appraisal format and earnest execution.
                  • In my opinion you can not  satisfy ML3-PA's-goals (how well institutionalized) without ML2 (what)
                  Please comment on proper alignment of my thought process?

                  I always say that a model is means to achieve improvement (even excellence) and the model is not the destination.

                  By  deciding  to go CMMI way , you got some directions for your journey (e.g. L5 which is optimizing a indication of continual process)

                  Regards Goyal


                  From: cmmi_process_ improvement@ yahoogroups. com [mailto:cmmi_process_improvement @yahoogroups. com] On Behalf Of Orhan KALAYCI
                  Sent: Monday, April 30, 2007 9:59 AM
                  To: cmmi_process_ improvement@ yahoogroups. com
                  Subject: {Disarmed} Re: [CMMi Process Improvement] Skipping ML2

                  Dear Goyal,

                  This is a good question.  And In fact, this question explains why we are (or I am alone) debating about it.

                  I am talking about this subject over and over because the importance of not skipping levels is widely over looked.  Going from ML1 to ML3 directly is counterproductive.  I believe there are a lot of companies at ML1 jump ML3.  Although these companies certified as ML3, their productivity do not increase even decrease. Then it looks like CMMI is counterproductive.  What is worse, people do not know that this counter productivity is a result of skipping levels but they think CMMI makes these companies less productive. 

                  May the peace be upon you,
                  Orhan
                  www.xpi.ca
                  www.nitelik. net


                  2007/4/30, R C Goyal < rcgoyal@gmail. com>:

                  Actually when an organization is being appraised for higher level the appraisal is in any case being done for satisfying goals of all PAs even for lower levels. Then why this subject of skipping is being debated at all.



                  On 4/29/07, Orhan KALAYCI < orhan.kalayci@ xpi.ca> wrote:

                  Hello,

                  I have difficulty in understanding why the most experienced CMMI gurus do not support my proposal:  Let us not encourage organizations to skip ML2! 

                  This is counter productive!  Do you really believe in it?  Do you support not skipping MLs especially ML2?  (We see nobody targeting directly ML5, it is ML2, the most skipped)

                  My questions:
                  -  What value do you think ML2 provides organizations with?
                  -  Do you think skipping ML2 is really counterproductive?  What kind of counter-productiven ess does it create?
                  -  How can an organization understand it is ready to start ML3 activities?  Not skipping ML2?

                  May the peace be upon you,
                  Orhan
                  www.xpi.ca
                  www.nitelik. net

                  You know my analogy:  "Before federal and state laws (standard processes) take over, we need to let sheriffs (PMs) to rule their counties (Projects) for a while.  We know, this approach has been proved to be successful in USA.  When ML2 tried to be skipped in Soviet Unions, where they rushed to standardize everything to catch the economics in scale.  We now know this approach (skipping ML2) is not successful, on the opposite it is really counterproductive.  At first sight, it looks like standards should provide productivity, because everybody will work in the same way, synchronization,  harmony, etc will be established.  However, if you rush to standardize too quick.  It results in just the opposite.  As I keep saying, life is full of contradictions."




                  --
                  Thanks & Regards R C Goyal
                  Management Committee  Member,  PMI & CSI Mumbai Chapter
                  BE AN AMBASSADOR OF QUALITY
                  programs@pmimumbaic hapter.org, rcgoyal@gmail. com; rcgoyal@avrms. com;
                  Mob: 00919869463964; 00912267108110





                  --
                  Thanks & Regards R C Goyal
                  Management Committee  Member,  PMI & CSI Mumbai Chapter
                  BE AN AMBASSADOR OF QUALITY
                  programs@pmimumbaic hapter.org, rcgoyal@gmail. com; rcgoyal@avrms. com;
                  Mob: 00919869463964; 00912267108110

                • Orhan KALAYCI
                  Andre, I see your point. My proposal for this situation can be: 1) At this level (ML2 or CL2) make process descriptions really simple. Just verification
                  Message 8 of 30 , May 1, 2007
                    Andre,

                    I see your point. 

                    My proposal for this situation can be:
                    1)  At this level (ML2 or CL2) make process descriptions really simple.  Just verification points, inputs, outputs, etc.
                    2)  If projects (counties) are too small, create some program managers (sheriffs) over project managers of smaller projects. 
                    3)  Let program managers be free to develop their own processes for their set of smaller projects in their control (county)

                    May the peace be upon you,
                    Orhan
                    www.xpi.ca
                    www.nitelik.net


                    2007/5/1, Andre Heijstek <andre.heijstek@...>:

                    Well, let me add my two cents to the discussion.

                    I assume that when you say skipping maturity levels, you mean skipping the spirit of a maturity level. In particular, when you 'skip' ML2 and go for ML3 directly, you skip the idea of having each project having its own process definitions, but using an organization-wide process description directly.
                    This is the only definition of skipping that I can think of that makes sense. It is not possible to skip process areas, because when you perform a ML3 appraisal you have to appraise the ML2 PA's as well.

                    In principle I agree with the points that Peter has made about the value of performing the spirit of ML2 first, I still want to make some comments. I believe that this interpretation makes most sense in the context where the CMMI was developed: large organizations that run few, large projects. Employees typically work for one project at a time.
                    Now, what to do when your context is significantly different? Imagine a organization having many small projects with at most 3 or 4 persons working for each project, and where many employees work for multiple projects in parallel. These projects do not have the time to develop their own processes in the first place. And even if the would have the time, think about those poor employees who need to remember that from 9:00 – 12:00 they work in project A using this set of processes, and from 13:00 – 17:00 they work for project B with a different set of processes!

                    So, I suggest that, instead of making absolute claims about which approach is the correct one, you have to take context into account. If your context is significantly different from the one that the CMMI authors had in mind, you will have to interpret the model in creative ways in order to still achieve the benefits.
                    Given this suggestion, the question becomes, what did the CMMI authors assume about project and organizational context? Let me start the list (hoping that others will contribute) by what I think the assumptions were:

                    • large organizations
                    • contracting relationship, that is: one well defined customer (and not a whole market of customers like for example Microsoft has)
                    • few projects in comparison to the number of employees
                    • most people work for one project at a time
                    (Note that the points above lead to relatively powerful project managers, they control many people, big budgets.)
                    • strong hierarchical reporting relationships – if the boss gives his orders, the employees have to follow, as opposed to a consensus-driven culture (that we in the Netherlands are more used to)
                    • a plan-driven culture; if you do something big and important, it is always good to plan first

                    Regards, Andre.

                    On 01-05-2007 08:29, "Orhan KALAYCI" <orhan.kalayci@xpi.ca> wrote:


                     
                     

                    Hello group, how are you today?

                    Peter, thank you very much for the message.  I was expecting Heather will explain "why skipping ML2 is counterproductive" but yours is very good, too.  Still, I am expecting Heather and others to share their valuable comments.

                    If we can not develop enough best practices in ML2 (or CL2) there will not be much (or bests) to share as a standard in ML3 (or CL3).  On the other hand, if we rush for standards they will be just standards, not the best practices shared.  

                    Skipping levels (esp. ML2) is counter productive in both representations, continuous or staged.  

                    We should not think levels (in both representations) as if they are just an arithmetic set of PAs within which GGs, SGs, etc.  There is a spirit behind CMMI and levels (maturity or capability).  Do you remember how Rocky wins against the Soviet opponent (representing the soulless robot) ? (in one of the Rocky X? movies :)  With his spirit!  ( Thanks google:   http://www.imdb.com/title/tt0089927/ )

                    May the peace be upon you,
                    Orhan
                    www.xpi.ca <http://www.xpi.ca>
                    www.nitelik.net  <http://www.nitelik.net>


                    2007/4/30, Peter Leeson <qpitpjl@yahoo.co.uk>:

                     
                     

                    Hello Orhan
                      
                     
                      
                    ML2 process areas are areas that focus on discovery: what are we trying to achieve? What are the measurements and reports that matter to management? How can we best do this? What is satisfying our corporate quality requirements?...
                      
                     
                      
                    With minimum guidelines (policy explaining the needs, clear measurement needs, management overview, quality assurance), we can let every team, project, group determine how they would like to respond to the requirements of the company, the management and the customer. By implementing the controls and overviews required ( GP2.9 and GP2.10 - which are different), we can determine what are the practices, the processes, the tools that really work for this company, this environment, these people, these customers, these pressures. At ML3, we are looking at how we can share the best practices. If a number of practices are documented and imposed centrally, we will never know whether they are best practices or just reasonable practices. I strongly recommend to encourage differences at ML2, so that approaches can be measured and compared to identify those best practices.
                      
                     
                      
                    Most organizations that jump to ML3 end up imposing an approach that someone thought was good, better practices are banned as compliance is enforced.
                      
                     
                      
                    Peter Leeson
                      
                    www.qpit.ltd.uk <http://www.qpit.ltd.uk>
                      

                    Yahoo! Mail is the world's favourite email. Don't settle for less, sign up for your free account today <http://uk.rd.yahoo.com/evt=44106/*http://uk.docs.yahoo.com/mail/winter07.html> .

                     

                        

                     
                        


                  • Andre Heijstek
                    Orhan, Nice try, but: 1. Very simple process descriptions probably don¹t add any value. In that case everybody can still do whatever s/he likes, so that is
                    Message 9 of 30 , May 1, 2007
                      Re: [CMMi Process Improvement] Re:Skipping ML2 Orhan,

                      Nice try, but:
                      1. Very simple process descriptions probably don’t add any value. In that case everybody can still do whatever s/he likes, so that is equivalent to having no process at all.
                      2. Yes, of course, this is what is typically called a department
                      3. Yes, then we have department-wide (which is organisation-wide) processes.

                      Regards, Andre.


                      On 01-05-2007 19:15, "Orhan KALAYCI" <orhan.kalayci@...> wrote:


                       
                       

                      Andre,

                      I see your point.  

                      My proposal for this situation can be:
                      1)  At this level (ML2 or CL2) make process descriptions really simple.  Just verification points, inputs, outputs, etc.
                      2)  If projects (counties) are too small, create some program managers (sheriffs) over project managers of smaller projects.  
                      3)  Let program managers be free to develop their own processes for their set of smaller projects in their control (county)

                      May the peace be upon you,
                      Orhan
                      www.xpi.ca <http://www.xpi.ca>
                      www.nitelik.net <http://www.nitelik.net>


                      2007/5/1, Andre Heijstek <andre.heijstek@...>:

                       
                       

                      Well, let me add my two cents to the discussion.

                      I assume that when you say skipping maturity levels, you mean skipping the spirit of a maturity level. In particular, when you 'skip' ML2 and go for ML3 directly, you skip the idea of having each project having its own process definitions, but using an organization-wide process description directly.
                      This is the only definition of skipping that I can think of that makes sense. It is not possible to skip process areas, because when you perform a ML3 appraisal you have to appraise the ML2 PA's as well.

                      In principle I agree with the points that Peter has made about the value of performing the spirit of ML2 first, I still want to make some comments. I believe that this interpretation makes most sense in the context where the CMMI was developed: large organizations that run few, large projects. Employees typically work for one project at a time.
                      Now, what to do when your context is significantly different? Imagine a organization having many small projects with at most 3 or 4 persons working for each project, and where many employees work for multiple projects in parallel. These projects do not have the time to develop their own processes in the first place. And even if the would have the time, think about those poor employees who need to remember that from 9:00 – 12:00 they work in project A using this set of processes, and from 13:00 – 17:00 they work for project B with a different set of processes!

                      So, I suggest that, instead of making absolute claims about which approach is the correct one, you have to take context into account. If your context is significantly different from the one that the CMMI authors had in mind, you will have to interpret the model in creative ways in order to still achieve the benefits.
                      Given this suggestion, the question becomes, what did the CMMI authors assume about project and organizational context? Let me start the list (hoping that others will contribute) by what I think the assumptions were:
                      • large organizations
                      • contracting relationship, that is: one well defined customer (and not a whole market of customers like for example Microsoft has)
                      • few projects in comparison to the number of employees
                      • most people work for one project at a time
                      (Note that the points above lead to relatively powerful project managers, they control many people, big budgets.)
                      • strong hierarchical reporting relationships – if the boss gives his orders, the employees have to follow, as opposed to a consensus-driven culture (that we in the Netherlands are more used to)
                      • a plan-driven culture; if you do something big and important, it is always good to plan first

                      Regards, Andre.

                      On 01-05-2007 08:29, "Orhan KALAYCI" <orhan.kalayci@... <http://xpi.ca> > wrote:


                       
                       

                      Hello group, how are you today?

                      Peter, thank you very much for the message.  I was expecting Heather will explain "why skipping ML2 is counterproductive" but yours is very good, too.  Still, I am expecting Heather and others to share their valuable comments.

                      If we can not develop enough best practices in ML2 (or CL2) there will not be much (or bests) to share as a standard in ML3 (or CL3).  On the other hand, if we rush for standards they will be just standards, not the best practices shared.  

                      Skipping levels (esp. ML2) is counter productive in both representations, continuous or staged.  

                      We should not think levels (in both representations) as if they are just an arithmetic set of PAs within which GGs, SGs, etc.  There is a spirit behind CMMI and levels (maturity or capability).  Do you remember how Rocky wins against the Soviet opponent (representing the soulless robot) ? (in one of the Rocky X? movies :)  With his spirit!  ( Thanks google:  http://www.imdb.com/title/tt0089927/ )

                      May the peace be upon you,
                      Orhan
                      www.xpi.ca <http://www.xpi.ca>  <http://www.xpi.ca>  <http://www.xpi.ca>  
                      www.nitelik.net <http://www.nitelik.net>   <http://www.nitelik.net>


                      2007/4/30, Peter Leeson <qpitpjl@... <http://co.uk> >:

                       
                       

                      Hello Orhan
                        
                       
                        
                      ML2 process areas are areas that focus on discovery: what are we trying to achieve? What are the measurements and reports that matter to management? How can we best do this? What is satisfying our corporate quality requirements?...
                        
                       
                        
                      With minimum guidelines (policy explaining the needs, clear measurement needs, management overview, quality assurance), we can let every team, project, group determine how they would like to respond to the requirements of the company, the management and the customer. By implementing the controls and overviews required ( GP2.9 and GP2.10 - which are different), we can determine what are the practices, the processes, the tools that really work for this company, this environment, these people, these customers, these pressures. At ML3, we are looking at how we can share the best practices. If a number of practices are documented and imposed centrally, we will never know whether they are best practices or just reasonable practices. I strongly recommend to encourage differences at ML2, so that approaches can be measured and compared to identify those best practices.
                        
                       
                        
                      Most organizations that jump to ML3 end up imposing an approach that someone thought was good, better practices are banned as compliance is enforced.
                        
                       
                        
                      Peter Leeson
                        
                      www.qpit.ltd.uk <http://www.qpit.ltd.uk>  <http://www.qpit.ltd.uk>
                        

                      Yahoo! Mail is the world's favourite email. Don't settle for less, sign up for your free account today <http://uk.rd.yahoo.com/evt=44106/*http://uk.docs.yahoo.com/mail/winter07.html> .

                       

                          

                       
                          


                       
                          

                       
                          

                    • Orhan KALAYCI
                      Hello group, how are you? To be able discuss or debate on ML2, ML3 or CL2, CL3. Let me define what they mean: First of all, an ML is not equal to a set of PAs
                      Message 10 of 30 , May 2, 2007
                        Hello group, how are you?

                        To be able discuss or debate on ML2, ML3 or CL2, CL3.  Let me define what they mean:
                        First of all, an ML is not equal to a set of PAs i.e. ML <> {PAs}

                        CMMI has a spirit, so do each ML and CL. 

                        ML2 means "Project management" is in place including processes (tasks) for process areas of ML3 in the project plans. However, they all are at the level of CL2.  We can never think about a project plan without validation or verification tasks in it.  So at ML2, it is for sure there are processes representing VER and VAL PAs (at CL2).  However they are not tailored from set of organizational standard processes (CL3).  Let me continue, can you think of a real project plan without processes representing TS, PI, RD?  No, it is nonsense to have a project plan, without requirements analysis or developing a solution (design, coding, etc).  Is there a chance that we can deliver our product to customer without integrating product components?  At ML2, they (processes (tasks) for PAs of ML2 + ML3) all should be there but they are all at CL2.  Spirit of ML3 is not a set of PAs.  What makes an organization at ML3 (spirit) or a process at CL3 (spirit)  is GG3, i.e., standard processes and tailoring guideline.  Management of standard processes is OPD. 

                        OPF should belong ML2.  I will write a CR for that. 
                        In fact, I need to write another CR about making it "explicit" that ML2 = PAs of (ML2+ML3) should be at CL2.  This is now implicit in PP in WBS, Work packages, task descriptions, etc. 

                        ML2 and CL2 is about "living" the processes. 
                        ML3 and CL3 is about "writing" the processes.

                        So, do not start "writing" the processes (tasks) before "living" them.  If you do so, it is highly probable that you will write processes (tasks) unlivable.  And (kudos to Jeff) these processes unfortunately will not set you free, just the opposite.  They will not bring you joy and freedom but misery and slavery.  

                        May the peace and the freedom be upon you,
                        Orhan
                        www.xpi.ca
                        www.nitelik.net


                        2007/5/1, Patrick OToole < PACT.otoole@...>:

                         
                        Ah, I think I'm finally getting a handle on this discussion.  It seems to me there are two concurrent thought streams running through this dialogue (or at least through my head) - and it all comes down to the phrase "skipping maturity levels."
                         
                        One interpretation is "let's work on all the ML2 and ML3 process areas at the same time - doing ML2 first, and then working on ML3 afterwards is wimpy - real men do it all!"
                         
                        The other interpretation is "let's work on SOME of the ML3 things before we work on SOME of the ML2 things.  Afterall, we have project pain that would gain some relief from enhancing our Requirements Elicitation and/or Peer Review processes, and I know that these are ML3 things and that we're an ML1 organization, but still..."
                         
                        I would STRONGLY agree that typical ML1 organizations are NOT well-served to follow the first interpretation - let's do everything all at once.  Stabilize the project domain before you attack the organizational issues.  Get some credibility and momentum established before you start trying to conquer the world.  (I'll still contend that SOME ML1 organizations, who already have MANY of the ML2 process areas well established are probably in a position where they can start working on some of the ML3 process areas while they are finishing up the ML2 stuff, but that's not what I mean by a "typical ML1 organization.")
                         
                        Most of my earlier posts were focused on the second interpretation of "skipping maturity levels."  I am certainly NOT a purist when it comes to establishing infrastructure and fixing pain.  If an SEPG-like thing is going to help us get focus and momentum for process improvement, then let's dip into OPF and figure out what an SEPG is and what kinds of things they do.  If OPD can give us a bit of guidance on establishing life cycles, how to write process documentation, and establishing a measurement repository, then let's take a look at that as well.
                         
                        For pain relief - if improving the peer review process is going to eliminate some pain that we feel in unit, integration, and system testing, then I'm all for it.  If introducing JAD sessions is going to get us better requirements, then I'm all for that too. 
                         
                        I do not adhere to the philosophy that "you must ONLY do ML2 stuff before you work on any of the ML3 stuff" - and if THAT's what someone means by "skipping maturity levels" then they'd better be prepared for a vigorous debate. 
                         
                        We must also keep in mind, as I've mentioned before, that organizations may already have a bunch of good stuff in place when they first pick up the model.  They may already have a Project Management Office that has some good risk management practices and may be running IPM-like integrated programs.  They may have an independent test group that is doing their portion of VER extremely well.  They may have a separate training organization that is already outstanding at many of the OT practices.  I would NOT tell these organizations to "quit wasting their time improving these ML3 like practices because there are some gaps at ML2."
                         
                        One final thought - with 11 process areas, I would contend that ML3 is way too much for a typical ML2 organization to bite off all at once.  The typical ML2 organization might be better served to focus on 3-5 ML3 process areas first - and should probably first focus on those where they perceive the greatest pain reduction or additional benefit.  I'll contend that the typical ML1 organization would be far better served to focus on 2-3 ML2 process areas rather than forming 7 groups, which just HAPPEN to have the same names as each o the ML2 process areas!
                         
                        I've never been a big believer in trying to achieve maturity levels as quickly as possible.  I would much prefer that the organization focus on gaining the benefits for the projects in a sustainable way than gaining another certificate for the trophy case.  After all, certifications don't produce software - people and projects do.
                         
                        Regards,
                         
                        Pat
                         
                         
                         
                         
                        ----- Original Message -----
                        From: R C Goyal
                        Sent: Monday, April 30, 2007 10:54 PM
                        Subject: Re: {Disarmed} Re: [CMMi Process Improvement] Skipping ML2

                        Dear Orhan

                        Without entering into lengthy discussions may I mention:

                        • Measure about process being counterproductive is not easy. It may also be one's own view point depending upon one's own expectation
                        • Again I would say going from ML1 to ML3  does not mean skipping ML2, however it would also depend the meticulous gap-analysis and improvements for correctly satisfying goals of ML2-PAs and to great extent would depend on  quality of Class C (& B) appraisal format and earnest execution.
                        • In my opinion you can not  satisfy ML3-PA's-goals (how well institutionalized) without ML2 (what)
                        Please comment on proper alignment of my thought process?

                        I always say that a model is means to achieve improvement (even excellence) and the model is not the destination.

                        By  deciding  to go CMMI way , you got some directions for your journey (e.g. L5 which is optimizing a indication of continual process)

                        Regards Goyal


                        From: cmmi_process_improvement@yahoogroups.com [mailto:cmmi_ process_improvement@yahoogroups.com] On Behalf Of Orhan KALAYCI
                        Sent: Monday, April 30, 2007 9:59 AM
                        To: cmmi_process_improvement@yahoogroups.com
                        Subject: {Disarmed} Re: [CMMi Process Improvement] Skipping ML2

                        Dear Goyal,

                        This is a good question.  And In fact, this question explains why we are (or I am alone) debating about it.

                        I am talking about this subject over and over because the importance of not skipping levels is widely over looked.  Going from ML1 to ML3 directly is counterproductive.  I believe there are a lot of companies at ML1 jump ML3.  Although these companies certified as ML3, their productivity do not increase even decrease. Then it looks like CMMI is counterproductive.  What is worse, people do not know that this counter productivity is a result of skipping levels but they think CMMI makes these companies less productive. 

                        May the peace be upon you,
                        Orhan
                        www.xpi.ca
                        www.nitelik.net


                        2007/4/30, R C Goyal < rcgoyal@...>:

                        Actually when an organization is being appraised for higher level the appraisal is in any case being done for satisfying goals of all PAs even for lower levels. Then why this subject of skipping is being debated at all.



                        On 4/29/07, Orhan KALAYCI < orhan.kalayci@...> wrote:

                        Hello,

                        I have difficulty in understanding why the most experienced CMMI gurus do not support my proposal:  Let us not encourage organizations to skip ML2! 

                        This is counter productive!  Do you really believe in it?  Do you support not skipping MLs especially ML2?  (We see nobody targeting directly ML5, it is ML2, the most skipped)

                        My questions:
                        -  What value do you think ML2 provides organizations with?
                        -  Do you think skipping ML2 is really counterproductive?  What kind of counter-productiveness does it create?
                        -  How can an organization understand it is ready to start ML3 activities?  Not skipping ML2?

                        May the peace be upon you,
                        Orhan
                        www.xpi.ca
                        www.nitelik.net

                        You know my analogy:  "Before federal and state laws (standard processes) take over, we need to let sheriffs (PMs) to rule their counties (Projects) for a while.  We know, this approach has been proved to be successful in USA.  When ML2 tried to be skipped in Soviet Unions, where they rushed to standardize everything to catch the economics in scale.  We now know this approach (skipping ML2) is not successful, on the opposite it is really counterproductive.  At first sight, it looks like standards should provide productivity, because everybody will work in the same way, synchronization,  harmony, etc will be established.  However, if you rush to standardize too quick.  It results in just the opposite.  As I keep saying, life is full of contradictions."




                        --
                        Thanks & Regards R C Goyal
                        Management Committee  Member,  PMI & CSI Mumbai Chapter
                        BE AN AMBASSADOR OF QUALITY
                        programs@..., rcgoyal@...; rcgoyal@...;
                        Mob: 00919869463964; 00912267108110





                        --
                        Thanks & Regards R C Goyal
                        Management Committee  Member,  PMI & CSI Mumbai Chapter
                        BE AN AMBASSADOR OF QUALITY
                        programs@..., rcgoyal@...; rcgoyal@...;
                        Mob: 00919869463964; 00912267108110


                      • Patrick OToole
                        Orhan, If an ML1 organization already has documented processes for its software and test engineering activities, and these documented processes include ways to
                        Message 11 of 30 , May 2, 2007
                          
                           
                          Orhan,
                           
                          If an ML1 organization already has documented processes for its software and test engineering activities, and these documented processes include ways to tailor them to best meet their needs, should they abandon them until they completely implement all practices associated with ML2?
                           
                          Regards,
                           
                          Pat
                           
                           
                          ----- Original Message -----
                          Sent: Wednesday, May 02, 2007 1:29 PM
                          Subject: Re: {Disarmed} Re: [CMMi Process Improvement] Skipping ML2

                          Hello group, how are you?

                          To be able discuss or debate on ML2, ML3 or CL2, CL3.  Let me define what they mean:
                          First of all, an ML is not equal to a set of PAs i.e. ML <> {PAs}

                          CMMI has a spirit, so do each ML and CL. 

                          ML2 means "Project management" is in place including processes (tasks) for process areas of ML3 in the project plans. However, they all are at the level of CL2.  We can never think about a project plan without validation or verification tasks in it.  So at ML2, it is for sure there are processes representing VER and VAL PAs (at CL2).  However they are not tailored from set of organizational standard processes (CL3).  Let me continue, can you think of a real project plan without processes representing TS, PI, RD?  No, it is nonsense to have a project plan, without requirements analysis or developing a solution (design, coding, etc).  Is there a chance that we can deliver our product to customer without integrating product components?  At ML2, they (processes (tasks) for PAs of ML2 + ML3) all should be there but they are all at CL2.  Spirit of ML3 is not a set of PAs.  What makes an organization at ML3 (spirit) or a process at CL3 (spirit)  is GG3, i.e., standard processes and tailoring guideline.  Management of standard processes is OPD. 

                          OPF should belong ML2.  I will write a CR for that. 
                          In fact, I need to write another CR about making it "explicit" that ML2 = PAs of (ML2+ML3) should be at CL2.  This is now implicit in PP in WBS, Work packages, task descriptions, etc. 

                          ML2 and CL2 is about "living" the processes. 
                          ML3 and CL3 is about "writing" the processes.

                          So, do not start "writing" the processes (tasks) before "living" them.  If you do so, it is highly probable that you will write processes (tasks) unlivable.  And (kudos to Jeff) these processes unfortunately will not set you free, just the opposite.  They will not bring you joy and freedom but misery and slavery.  

                          May the peace and the freedom be upon you,
                          Orhan
                          www.xpi.ca
                          www.nitelik. net


                          2007/5/1, Patrick OToole < PACT.otoole@ att.net>:

                           
                          Ah, I think I'm finally getting a handle on this discussion.  It seems to me there are two concurrent thought streams running through this dialogue (or at least through my head) - and it all comes down to the phrase "skipping maturity levels."
                           
                          One interpretation is "let's work on all the ML2 and ML3 process areas at the same time - doing ML2 first, and then working on ML3 afterwards is wimpy - real men do it all!"
                           
                          The other interpretation is "let's work on SOME of the ML3 things before we work on SOME of the ML2 things.  Afterall, we have project pain that would gain some relief from enhancing our Requirements Elicitation and/or Peer Review processes, and I know that these are ML3 things and that we're an ML1 organization, but still..."
                           
                          I would STRONGLY agree that typical ML1 organizations are NOT well-served to follow the first interpretation - let's do everything all at once.  Stabilize the project domain before you attack the organizational issues.  Get some credibility and momentum established before you start trying to conquer the world.  (I'll still contend that SOME ML1 organizations, who already have MANY of the ML2 process areas well established are probably in a position where they can start working on some of the ML3 process areas while they are finishing up the ML2 stuff, but that's not what I mean by a "typical ML1 organization.")
                           
                          Most of my earlier posts were focused on the second interpretation of "skipping maturity levels."  I am certainly NOT a purist when it comes to establishing infrastructure and fixing pain.  If an SEPG-like thing is going to help us get focus and momentum for process improvement, then let's dip into OPF and figure out what an SEPG is and what kinds of things they do.  If OPD can give us a bit of guidance on establishing life cycles, how to write process documentation, and establishing a measurement repository, then let's take a look at that as well.
                           
                          For pain relief - if improving the peer review process is going to eliminate some pain that we feel in unit, integration, and system testing, then I'm all for it.  If introducing JAD sessions is going to get us better requirements, then I'm all for that too. 
                           
                          I do not adhere to the philosophy that "you must ONLY do ML2 stuff before you work on any of the ML3 stuff" - and if THAT's what someone means by "skipping maturity levels" then they'd better be prepared for a vigorous debate. 
                           
                          We must also keep in mind, as I've mentioned before, that organizations may already have a bunch of good stuff in place when they first pick up the model.  They may already have a Project Management Office that has some good risk management practices and may be running IPM-like integrated programs.  They may have an independent test group that is doing their portion of VER extremely well.  They may have a separate training organization that is already outstanding at many of the OT practices.  I would NOT tell these organizations to "quit wasting their time improving these ML3 like practices because there are some gaps at ML2."
                           
                          One final thought - with 11 process areas, I would contend that ML3 is way too much for a typical ML2 organization to bite off all at once.  The typical ML2 organization might be better served to focus on 3-5 ML3 process areas first - and should probably first focus on those where they perceive the greatest pain reduction or additional benefit.  I'll contend that the typical ML1 organization would be far better served to focus on 2-3 ML2 process areas rather than forming 7 groups, which just HAPPEN to have the same names as each o the ML2 process areas!
                           
                          I've never been a big believer in trying to achieve maturity levels as quickly as possible.  I would much prefer that the organization focus on gaining the benefits for the projects in a sustainable way than gaining another certificate for the trophy case.  After all, certifications don't produce software - people and projects do.
                           
                          Regards,
                           
                          Pat
                           
                           
                           
                           
                          ----- Original Message -----
                          From: R C Goyal
                          Sent: Monday, April 30, 2007 10:54 PM
                          Subject: Re: {Disarmed} Re: [CMMi Process Improvement] Skipping ML2

                          Dear Orhan

                          Without entering into lengthy discussions may I mention:

                          • Measure about process being counterproductive is not easy. It may also be one's own view point depending upon one's own expectation
                          • Again I would say going from ML1 to ML3  does not mean skipping ML2, however it would also depend the meticulous gap-analysis and improvements for correctly satisfying goals of ML2-PAs and to great extent would depend on  quality of Class C (& B) appraisal format and earnest execution.
                          • In my opinion you can not  satisfy ML3-PA's-goals (how well institutionalized) without ML2 (what)
                          Please comment on proper alignment of my thought process?

                          I always say that a model is means to achieve improvement (even excellence) and the model is not the destination.

                          By  deciding  to go CMMI way , you got some directions for your journey (e.g. L5 which is optimizing a indication of continual process)

                          Regards Goyal


                          From: cmmi_process_ improvement@ yahoogroups. com [mailto:cmmi_ process_improvement @yahoogroups. com] On Behalf Of Orhan KALAYCI
                          Sent: Monday, April 30, 2007 9:59 AM
                          To: cmmi_process_ improvement@ yahoogroups. com
                          Subject: {Disarmed} Re: [CMMi Process Improvement] Skipping ML2

                          Dear Goyal,

                          This is a good question.  And In fact, this question explains why we are (or I am alone) debating about it.

                          I am talking about this subject over and over because the importance of not skipping levels is widely over looked.  Going from ML1 to ML3 directly is counterproductive.  I believe there are a lot of companies at ML1 jump ML3.  Although these companies certified as ML3, their productivity do not increase even decrease. Then it looks like CMMI is counterproductive.  What is worse, people do not know that this counter productivity is a result of skipping levels but they think CMMI makes these companies less productive. 

                          May the peace be upon you,
                          Orhan
                          www.xpi.ca
                          www.nitelik. net


                          2007/4/30, R C Goyal < rcgoyal@gmail. com>:

                          Actually when an organization is being appraised for higher level the appraisal is in any case being done for satisfying goals of all PAs even for lower levels. Then why this subject of skipping is being debated at all.



                          On 4/29/07, Orhan KALAYCI < orhan.kalayci@ xpi.ca> wrote:

                          Hello,

                          I have difficulty in understanding why the most experienced CMMI gurus do not support my proposal:  Let us not encourage organizations to skip ML2! 

                          This is counter productive!  Do you really believe in it?  Do you support not skipping MLs especially ML2?  (We see nobody targeting directly ML5, it is ML2, the most skipped)

                          My questions:
                          -  What value do you think ML2 provides organizations with?
                          -  Do you think skipping ML2 is really counterproductive?  What kind of counter-productiven ess does it create?
                          -  How can an organization understand it is ready to start ML3 activities?  Not skipping ML2?

                          May the peace be upon you,
                          Orhan
                          www.xpi.ca
                          www.nitelik. net

                          You know my analogy:  "Before federal and state laws (standard processes) take over, we need to let sheriffs (PMs) to rule their counties (Projects) for a while.  We know, this approach has been proved to be successful in USA.  When ML2 tried to be skipped in Soviet Unions, where they rushed to standardize everything to catch the economics in scale.  We now know this approach (skipping ML2) is not successful, on the opposite it is really counterproductive.  At first sight, it looks like standards should provide productivity, because everybody will work in the same way, synchronization,  harmony, etc will be established.  However, if you rush to standardize too quick.  It results in just the opposite.  As I keep saying, life is full of contradictions."




                          --
                          Thanks & Regards R C Goyal
                          Management Committee  Member,  PMI & CSI Mumbai Chapter
                          BE AN AMBASSADOR OF QUALITY
                          programs@pmimumbaic hapter.org, rcgoyal@gmail. com; rcgoyal@avrms. com;
                          Mob: 00919869463964; 00912267108110





                          --
                          Thanks & Regards R C Goyal
                          Management Committee  Member,  PMI & CSI Mumbai Chapter
                          BE AN AMBASSADOR OF QUALITY
                          programs@pmimumbaic hapter.org, rcgoyal@gmail. com; rcgoyal@avrms. com;
                          Mob: 00919869463964; 00912267108110


                        • Patrick OToole
                          Everyone, Here s another problem that I m having with the don t do things at ML3 until you ve got ML2 nailed position. Let s say an organization has
                          Message 12 of 30 , May 2, 2007
                            
                             
                            Everyone,
                             
                            Here's another "problem" that I'm having with the "don't do things at ML3 until you've got ML2 nailed" position.  Let's say an organization has staffed its process improvement project such that it can focus on improving 7 things concurrently.  Let's say they initially form one team to work on REQM, another to work on PP, another to work on PMC, etc.
                             
                            But, in this organization, project management is already in pretty good shape, while MA and PPQA are pretty much starting from scratch.  In such a case, the PP and PMC teams may wrap up pretty quickly - within a month or two they may fill the couple of gaps, train the other PMs to adopt the new practices in a value-added way and they're done.
                             
                            In the meantime, the MA and PPQA teams are trying to understand what the associated practices are all about, they are off getting trained on their respective process areas, and have put together a one-year plan to become "model compliant."
                             
                            Since the organization is already staffed to run 7 concurrent improvement teams, they COULD re-allocate all the PP and PMC folks to assist with MA and PPQA, and they'd wind up with twice as many people in a group grope.  The PP and PMC teams would now need to be trained and would be playing 2 months of catch up.  This MAY reduce the 12 month schedule by a month or two - or it may not.
                             
                            Alternatively, the PP and PMC teams could be re-directed to work on RSKM and IPM, maybe targeting CL1 or CL2 in these two process areas.  After all, this is simply an extension of the project management work they have already performed AND there may be a pretty solid foundation in some of these practices as well.
                             
                            Would the organization be WRONG if they decided to go with the second option?
                             
                            And, if they get IPM and RSKM to CL2 within 6 months, would they be wrong to then start working on PP and PMC GP3.1 and GP3.2?  Is there anything wrong with trying to standardize the practices and processes that were pretty much deeply institutionalized anyway?
                             
                            As much as I agree with the basic principle that we "shouldn't skip maturity levels," I'm merely suggesting that it is hardly ever that simple, and that we really need to look at each unique situation to determine what's best for THAT organization.  Just as there is no one way to implement any given CMMI practice, and we encourage organizations to implement the practices in a way that best fits their needs, shouldn't we be somewhat flexible in our statements about the ORDER in which they implement the practices?
                             
                            And on it goes...
                             
                            Regards,
                             
                            Pat
                             
                             
                             
                             
                            ----- Original Message -----
                            Sent: Wednesday, May 02, 2007 1:29 PM
                            Subject: Re: {Disarmed} Re: [CMMi Process Improvement] Skipping ML2

                            Hello group, how are you?

                            To be able discuss or debate on ML2, ML3 or CL2, CL3.  Let me define what they mean:
                            First of all, an ML is not equal to a set of PAs i.e. ML <> {PAs}

                            CMMI has a spirit, so do each ML and CL. 

                            ML2 means "Project management" is in place including processes (tasks) for process areas of ML3 in the project plans. However, they all are at the level of CL2.  We can never think about a project plan without validation or verification tasks in it.  So at ML2, it is for sure there are processes representing VER and VAL PAs (at CL2).  However they are not tailored from set of organizational standard processes (CL3).  Let me continue, can you think of a real project plan without processes representing TS, PI, RD?  No, it is nonsense to have a project plan, without requirements analysis or developing a solution (design, coding, etc).  Is there a chance that we can deliver our product to customer without integrating product components?  At ML2, they (processes (tasks) for PAs of ML2 + ML3) all should be there but they are all at CL2.  Spirit of ML3 is not a set of PAs.  What makes an organization at ML3 (spirit) or a process at CL3 (spirit)  is GG3, i.e., standard processes and tailoring guideline.  Management of standard processes is OPD. 

                            OPF should belong ML2.  I will write a CR for that. 
                            In fact, I need to write another CR about making it "explicit" that ML2 = PAs of (ML2+ML3) should be at CL2.  This is now implicit in PP in WBS, Work packages, task descriptions, etc. 

                            ML2 and CL2 is about "living" the processes. 
                            ML3 and CL3 is about "writing" the processes.

                            So, do not start "writing" the processes (tasks) before "living" them.  If you do so, it is highly probable that you will write processes (tasks) unlivable.  And (kudos to Jeff) these processes unfortunately will not set you free, just the opposite.  They will not bring you joy and freedom but misery and slavery.  

                            May the peace and the freedom be upon you,
                            Orhan
                            www.xpi.ca
                            www.nitelik. net


                            2007/5/1, Patrick OToole < PACT.otoole@ att.net>:

                             
                            Ah, I think I'm finally getting a handle on this discussion.  It seems to me there are two concurrent thought streams running through this dialogue (or at least through my head) - and it all comes down to the phrase "skipping maturity levels."
                             
                            One interpretation is "let's work on all the ML2 and ML3 process areas at the same time - doing ML2 first, and then working on ML3 afterwards is wimpy - real men do it all!"
                             
                            The other interpretation is "let's work on SOME of the ML3 things before we work on SOME of the ML2 things.  Afterall, we have project pain that would gain some relief from enhancing our Requirements Elicitation and/or Peer Review processes, and I know that these are ML3 things and that we're an ML1 organization, but still..."
                             
                            I would STRONGLY agree that typical ML1 organizations are NOT well-served to follow the first interpretation - let's do everything all at once.  Stabilize the project domain before you attack the organizational issues.  Get some credibility and momentum established before you start trying to conquer the world.  (I'll still contend that SOME ML1 organizations, who already have MANY of the ML2 process areas well established are probably in a position where they can start working on some of the ML3 process areas while they are finishing up the ML2 stuff, but that's not what I mean by a "typical ML1 organization.")
                             
                            Most of my earlier posts were focused on the second interpretation of "skipping maturity levels."  I am certainly NOT a purist when it comes to establishing infrastructure and fixing pain.  If an SEPG-like thing is going to help us get focus and momentum for process improvement, then let's dip into OPF and figure out what an SEPG is and what kinds of things they do.  If OPD can give us a bit of guidance on establishing life cycles, how to write process documentation, and establishing a measurement repository, then let's take a look at that as well.
                             
                            For pain relief - if improving the peer review process is going to eliminate some pain that we feel in unit, integration, and system testing, then I'm all for it.  If introducing JAD sessions is going to get us better requirements, then I'm all for that too. 
                             
                            I do not adhere to the philosophy that "you must ONLY do ML2 stuff before you work on any of the ML3 stuff" - and if THAT's what someone means by "skipping maturity levels" then they'd better be prepared for a vigorous debate. 
                             
                            We must also keep in mind, as I've mentioned before, that organizations may already have a bunch of good stuff in place when they first pick up the model.  They may already have a Project Management Office that has some good risk management practices and may be running IPM-like integrated programs.  They may have an independent test group that is doing their portion of VER extremely well.  They may have a separate training organization that is already outstanding at many of the OT practices.  I would NOT tell these organizations to "quit wasting their time improving these ML3 like practices because there are some gaps at ML2."
                             
                            One final thought - with 11 process areas, I would contend that ML3 is way too much for a typical ML2 organization to bite off all at once.  The typical ML2 organization might be better served to focus on 3-5 ML3 process areas first - and should probably first focus on those where they perceive the greatest pain reduction or additional benefit.  I'll contend that the typical ML1 organization would be far better served to focus on 2-3 ML2 process areas rather than forming 7 groups, which just HAPPEN to have the same names as each o the ML2 process areas!
                             
                            I've never been a big believer in trying to achieve maturity levels as quickly as possible.  I would much prefer that the organization focus on gaining the benefits for the projects in a sustainable way than gaining another certificate for the trophy case.  After all, certifications don't produce software - people and projects do.
                             
                            Regards,
                             
                            Pat
                             
                             
                             
                             
                            ----- Original Message -----
                            From: R C Goyal
                            Sent: Monday, April 30, 2007 10:54 PM
                            Subject: Re: {Disarmed} Re: [CMMi Process Improvement] Skipping ML2

                            Dear Orhan

                            Without entering into lengthy discussions may I mention:

                            • Measure about process being counterproductive is not easy. It may also be one's own view point depending upon one's own expectation
                            • Again I would say going from ML1 to ML3  does not mean skipping ML2, however it would also depend the meticulous gap-analysis and improvements for correctly satisfying goals of ML2-PAs and to great extent would depend on  quality of Class C (& B) appraisal format and earnest execution.
                            • In my opinion you can not  satisfy ML3-PA's-goals (how well institutionalized) without ML2 (what)
                            Please comment on proper alignment of my thought process?

                            I always say that a model is means to achieve improvement (even excellence) and the model is not the destination.

                            By  deciding  to go CMMI way , you got some directions for your journey (e.g. L5 which is optimizing a indication of continual process)

                            Regards Goyal


                            From: cmmi_process_ improvement@ yahoogroups. com [mailto:cmmi_ process_improvement @yahoogroups. com] On Behalf Of Orhan KALAYCI
                            Sent: Monday, April 30, 2007 9:59 AM
                            To: cmmi_process_ improvement@ yahoogroups. com
                            Subject: {Disarmed} Re: [CMMi Process Improvement] Skipping ML2

                            Dear Goyal,

                            This is a good question.  And In fact, this question explains why we are (or I am alone) debating about it.

                            I am talking about this subject over and over because the importance of not skipping levels is widely over looked.  Going from ML1 to ML3 directly is counterproductive.  I believe there are a lot of companies at ML1 jump ML3.  Although these companies certified as ML3, their productivity do not increase even decrease. Then it looks like CMMI is counterproductive.  What is worse, people do not know that this counter productivity is a result of skipping levels but they think CMMI makes these companies less productive. 

                            May the peace be upon you,
                            Orhan
                            www.xpi.ca
                            www.nitelik. net


                            2007/4/30, R C Goyal < rcgoyal@gmail. com>:

                            Actually when an organization is being appraised for higher level the appraisal is in any case being done for satisfying goals of all PAs even for lower levels. Then why this subject of skipping is being debated at all.



                            On 4/29/07, Orhan KALAYCI < orhan.kalayci@ xpi.ca> wrote:

                            Hello,

                            I have difficulty in understanding why the most experienced CMMI gurus do not support my proposal:  Let us not encourage organizations to skip ML2! 

                            This is counter productive!  Do you really believe in it?  Do you support not skipping MLs especially ML2?  (We see nobody targeting directly ML5, it is ML2, the most skipped)

                            My questions:
                            -  What value do you think ML2 provides organizations with?
                            -  Do you think skipping ML2 is really counterproductive?  What kind of counter-productiven ess does it create?
                            -  How can an organization understand it is ready to start ML3 activities?  Not skipping ML2?

                            May the peace be upon you,
                            Orhan
                            www.xpi.ca
                            www.nitelik. net

                            You know my analogy:  "Before federal and state laws (standard processes) take over, we need to let sheriffs (PMs) to rule their counties (Projects) for a while.  We know, this approach has been proved to be successful in USA.  When ML2 tried to be skipped in Soviet Unions, where they rushed to standardize everything to catch the economics in scale.  We now know this approach (skipping ML2) is not successful, on the opposite it is really counterproductive.  At first sight, it looks like standards should provide productivity, because everybody will work in the same way, synchronization,  harmony, etc will be established.  However, if you rush to standardize too quick.  It results in just the opposite.  As I keep saying, life is full of contradictions."




                            --
                            Thanks & Regards R C Goyal
                            Management Committee  Member,  PMI & CSI Mumbai Chapter
                            BE AN AMBASSADOR OF QUALITY
                            programs@pmimumbaic hapter.org, rcgoyal@gmail. com; rcgoyal@avrms. com;
                            Mob: 00919869463964; 00912267108110





                            --
                            Thanks & Regards R C Goyal
                            Management Committee  Member,  PMI & CSI Mumbai Chapter
                            BE AN AMBASSADOR OF QUALITY
                            programs@pmimumbaic hapter.org, rcgoyal@gmail. com; rcgoyal@avrms. com;
                            Mob: 00919869463964; 00912267108110


                          • R C Goyal
                            Dear Orhan One also need to know the context i.e. are we talking about Stagged representation or Continuous, because CL (capability level) concept is
                            Message 13 of 30 , May 2, 2007
                              Dear Orhan

                              One also need to know the context i.e. are we talking about Stagged representation or Continuous, because CL (capability level) concept is  considered in Continuous  representation.

                              In case of Continuous representation there no ML-concept for organization as whole, because  each PA  may be at different CL (which is maturity level of that PA).

                              Please comment on my understanding.

                              Regards Goyal

                              On 5/2/07, Orhan KALAYCI <orhan.kalayci@... > wrote:

                              Hello group, how are you?

                              To be able discuss or debate on ML2, ML3 or CL2, CL3.  Let me define what they mean:
                              First of all, an ML is not equal to a set of PAs i.e. ML <> {PAs}

                              CMMI has a spirit, so do each ML and CL. 

                              ML2 means "Project management" is in place including processes (tasks) for process areas of ML3 in the project plans. However, they all are at the level of CL2.  We can never think about a project plan without validation or verification tasks in it.  So at ML2, it is for sure there are processes representing VER and VAL PAs (at CL2).  However they are not tailored from set of organizational standard processes (CL3).  Let me continue, can you think of a real project plan without processes representing TS, PI, RD?  No, it is nonsense to have a project plan, without requirements analysis or developing a solution (design, coding, etc).  Is there a chance that we can deliver our product to customer without integrating product components?  At ML2, they (processes (tasks) for PAs of ML2 + ML3) all should be there but they are all at CL2.  Spirit of ML3 is not a set of PAs.  What makes an organization at ML3 (spirit) or a process at CL3 (spirit)  is GG3, i.e., standard processes and tailoring guideline.  Management of standard processes is OPD. 

                              OPF should belong ML2.  I will write a CR for that. 
                              In fact, I need to write another CR about making it "explicit" that ML2 = PAs of (ML2+ML3) should be at CL2.  This is now implicit in PP in WBS, Work packages, task descriptions, etc. 

                              ML2 and CL2 is about "living" the processes. 
                              ML3 and CL3 is about "writing" the processes.

                              So, do not start "writing" the processes (tasks) before "living" them.  If you do so, it is highly probable that you will write processes (tasks) unlivable.  And (kudos to Jeff) these processes unfortunately will not set you free, just the opposite.  They will not bring you joy and freedom but misery and slavery.  

                              May the peace and the freedom be upon you,
                              Orhan
                              www.xpi.ca
                              www.nitelik.net


                              2007/5/1, Patrick OToole < PACT.otoole@...>:

                               
                              Ah, I think I'm finally getting a handle on this discussion.  It seems to me there are two concurrent thought streams running through this dialogue (or at least through my head) - and it all comes down to the phrase "skipping maturity levels."
                               
                              One interpretation is "let's work on all the ML2 and ML3 process areas at the same time - doing ML2 first, and then working on ML3 afterwards is wimpy - real men do it all!"
                               
                              The other interpretation is "let's work on SOME of the ML3 things before we work on SOME of the ML2 things.  Afterall, we have project pain that would gain some relief from enhancing our Requirements Elicitation and/or Peer Review processes, and I know that these are ML3 things and that we're an ML1 organization, but still..."
                               
                              I would STRONGLY agree that typical ML1 organizations are NOT well-served to follow the first interpretation - let's do everything all at once.  Stabilize the project domain before you attack the organizational issues.  Get some credibility and momentum established before you start trying to conquer the world.  (I'll still contend that SOME ML1 organizations, who already have MANY of the ML2 process areas well established are probably in a position where they can start working on some of the ML3 process areas while they are finishing up the ML2 stuff, but that's not what I mean by a "typical ML1 organization.")
                               
                              Most of my earlier posts were focused on the second interpretation of "skipping maturity levels."  I am certainly NOT a purist when it comes to establishing infrastructure and fixing pain.  If an SEPG-like thing is going to help us get focus and momentum for process improvement, then let's dip into OPF and figure out what an SEPG is and what kinds of things they do.  If OPD can give us a bit of guidance on establishing life cycles, how to write process documentation, and establishing a measurement repository, then let's take a look at that as well.
                               
                              For pain relief - if improving the peer review process is going to eliminate some pain that we feel in unit, integration, and system testing, then I'm all for it.  If introducing JAD sessions is going to get us better requirements, then I'm all for that too. 
                               
                              I do not adhere to the philosophy that "you must ONLY do ML2 stuff before you work on any of the ML3 stuff" - and if THAT's what someone means by "skipping maturity levels" then they'd better be prepared for a vigorous debate. 
                               
                              We must also keep in mind, as I've mentioned before, that organizations may already have a bunch of good stuff in place when they first pick up the model.  They may already have a Project Management Office that has some good risk management practices and may be running IPM-like integrated programs.  They may have an independent test group that is doing their portion of VER extremely well.  They may have a separate training organization that is already outstanding at many of the OT practices.  I would NOT tell these organizations to "quit wasting their time improving these ML3 like practices because there are some gaps at ML2."
                               
                              One final thought - with 11 process areas, I would contend that ML3 is way too much for a typical ML2 organization to bite off all at once.  The typical ML2 organization might be better served to focus on 3-5 ML3 process areas first - and should probably first focus on those where they perceive the greatest pain reduction or additional benefit.  I'll contend that the typical ML1 organization would be far better served to focus on 2-3 ML2 process areas rather than forming 7 groups, which just HAPPEN to have the same names as each o the ML2 process areas!
                               
                              I've never been a big believer in trying to achieve maturity levels as quickly as possible.  I would much prefer that the organization focus on gaining the benefits for the projects in a sustainable way than gaining another certificate for the trophy case.  After all, certifications don't produce software - people and projects do.
                               
                              Regards,
                               
                              Pat
                               
                               
                               
                               
                              ----- Original Message -----
                              From: R C Goyal
                              Sent: Monday, April 30, 2007 10:54 PM
                              Subject: Re: {Disarmed} Re: [CMMi Process Improvement] Skipping ML2

                              Dear Orhan

                              Without entering into lengthy discussions may I mention:

                              • Measure about process being counterproductive is not easy. It may also be one's own view point depending upon one's own expectation
                              • Again I would say going from ML1 to ML3  does not mean skipping ML2, however it would also depend the meticulous gap-analysis and improvements for correctly satisfying goals of ML2-PAs and to great extent would depend on  quality of Class C (& B) appraisal format and earnest execution.
                              • In my opinion you can not  satisfy ML3-PA's-goals (how well institutionalized) without ML2 (what)
                              Please comment on proper alignment of my thought process?

                              I always say that a model is means to achieve improvement (even excellence) and the model is not the destination.

                              By  deciding  to go CMMI way , you got some directions for your journey (e.g. L5 which is optimizing a indication of continual process)

                              Regards Goyal


                              From: cmmi_process_improvement@yahoogroups.com [mailto:cmmi_process_improvement@yahoogroups.com] On Behalf Of Orhan KALAYCI
                              Sent: Monday, April 30, 2007 9:59 AM
                              To: cmmi_process_improvement@yahoogroups.com
                              Subject: {Disarmed} Re: [CMMi Process Improvement] Skipping ML2

                              Dear Goyal,

                              This is a good question.  And In fact, this question explains why we are (or I am alone) debating about it.

                              I am talking about this subject over and over because the importance of not skipping levels is widely over looked.  Going from ML1 to ML3 directly is counterproductive.  I believe there are a lot of companies at ML1 jump ML3.  Although these companies certified as ML3, their productivity do not increase even decrease. Then it looks like CMMI is counterproductive.  What is worse, people do not know that this counter productivity is a result of skipping levels but they think CMMI makes these companies less productive. 

                              May the peace be upon you,
                              Orhan
                              www.xpi.ca
                              www.nitelik.net


                              2007/4/30, R C Goyal < rcgoyal@...>:

                              Actually when an organization is being appraised for higher level the appraisal is in any case being done for satisfying goals of all PAs even for lower levels. Then why this subject of skipping is being debated at all.



                              On 4/29/07, Orhan KALAYCI < orhan.kalayci@...> wrote:

                              Hello,

                              I have difficulty in understanding why the most experienced CMMI gurus do not support my proposal:  Let us not encourage organizations to skip ML2! 

                              This is counter productive!  Do you really believe in it?  Do you support not skipping MLs especially ML2?  (We see nobody targeting directly ML5, it is ML2, the most skipped)

                              My questions:
                              -  What value do you think ML2 provides organizations with?
                              -  Do you think skipping ML2 is really counterproductive?  What kind of counter-productiveness does it create?
                              -  How can an organization understand it is ready to start ML3 activities?  Not skipping ML2?

                              May the peace be upon you,
                              Orhan
                              www.xpi.ca
                              www.nitelik.net

                              You know my analogy:  "Before federal and state laws (standard processes) take over, we need to let sheriffs (PMs) to rule their counties (Projects) for a while.  We know, this approach has been proved to be successful in USA.  When ML2 tried to be skipped in Soviet Unions, where they rushed to standardize everything to catch the economics in scale.  We now know this approach (skipping ML2) is not successful, on the opposite it is really counterproductive.  At first sight, it looks like standards should provide productivity, because everybody will work in the same way, synchronization,  harmony, etc will be established.  However, if you rush to standardize too quick.  It results in just the opposite.  As I keep saying, life is full of contradictions."




                              --
                              Thanks & Regards R C Goyal
                              Management Committee  Member,  PMI & CSI Mumbai Chapter
                              BE AN AMBASSADOR OF QUALITY
                              programs@..., rcgoyal@...; rcgoyal@...;
                              Mob: 00919869463964; 00912267108110





                              --
                              Thanks & Regards R C Goyal
                              Management Committee  Member,  PMI & CSI Mumbai Chapter
                              BE AN AMBASSADOR OF QUALITY
                              programs@..., rcgoyal@...; rcgoyal@...;
                              Mob: 00919869463964; 00912267108110





                              --
                              Thanks & Regards R C Goyal
                              Management Committee  Member,  PMI & CSI Mumbai Chapter
                              BE AN AMBASSADOR OF QUALITY
                              programs@..., rcgoyal@...; rcgoyal@...;
                              Mob: 00919869463964; 00912267108110
                            • Orhan KALAYCI
                              Hello Goyal, how are you? From spirit point of view, there is no difference between CLs and MLs. Their spirits are the same. In fact, in the spirit theory,
                              Message 14 of 30 , May 3, 2007
                                Hello Goyal, how are you?

                                From spirit point of view, there is no difference between CLs and MLs.  Their spirits are the same.  In fact, in the spirit theory, all the spirits are coming from the same source, you can name it according to your religion or belief.  If you lose the spirit of anything you can not be successful however you are clever, bright, smart, etc.  Just like the most powerful computer is a computer at the end.  Highest spiritual activity of human being is innovation, that is why, only innovative nations or organizations may live longer.  And only free souls and humans may be innovative (ML5)  As Jeff said, processes will set us free.  And as Orhan (me) said,  livable processes that will set us really free can only be defined by the people who had lived the processes long enough in the (same) organization. 

                                ML and CL are not from different contexts.  As CMMI correctly named them they are different representations of the same concept/context.  (Similar to different views in DBs)

                                In fact, spirit of both concept can be summarized as follows:

                                CL0 - The process does not exists at all. An LA may go to an organization and it is found that the processes does not even satisfy the SPs.  In other words, there is no process.  There is no chance where an LA goes and see there is no organization at all or at least in this case there is no need to write appraisal report.  That is why there is no ML0. :)

                                CL1/ML1 - Organization or process just exists.  It has a name, everybody knows it.  But you know, it is not always consistent.  Different times, different moods. 

                                CL2/ML2 - Ok.  It is now consistent.  Under stress it is not to the first to be forgotten.  Although everybody has his or her own interpretation, it is operational all the time. 

                                CL3/ML3 - Better now.  It is consistently performed and more then that everybody do it in the same way.  People have common understanding and common definition for the process(es). 

                                CL4/ML4 - Now, we are ready to measure it.  Because there is no risk of adding oranges and apples together, any more.  In the past, some may point out an apple ( e.g. SQA) and call it orange (e.g. test) in ML1/CL1. Now we now are sure when someone call something orange (e.g. test)  it is really orange (e.g. test).  We now count these orange make some arithmetics and statistical analysis such as means, deviations, upper and lower limits, etc.

                                CL5/ML5 - Improvement now is not a feeling, it is a fact.  When we change a process we can measure the results and make sure it is a real improvement.  You know there are some counter productive movement in lower levels that some may debate they are good and others may claim, no just the opposite. 

                                May the peace and the freedom and the spirit and the innovation be upon you,
                                Orhan
                                www.xpi.ca
                                www.nitelik.net


                                2007/5/3, R C Goyal <rcgoyal@...>:

                                Dear Orhan

                                One also need to know the context i.e. are we talking about Stagged representation or Continuous, because CL (capability level) concept is  considered in Continuous  representation.

                                In case of Continuous representation there no ML-concept for organization as whole, because  each PA  may be at different CL (which is maturity level of that PA).

                                Please comment on my understanding.

                                Regards Goyal



                                On 5/2/07, Orhan KALAYCI < orhan.kalayci@... > wrote:

                                Hello group, how are you?

                                To be able discuss or debate on ML2, ML3 or CL2, CL3.  Let me define what they mean:
                                First of all, an ML is not equal to a set of PAs i.e. ML <> {PAs}

                                CMMI has a spirit, so do each ML and CL. 

                                ML2 means "Project management" is in place including processes (tasks) for process areas of ML3 in the project plans. However, they all are at the level of CL2.  We can never think about a project plan without validation or verification tasks in it.  So at ML2, it is for sure there are processes representing VER and VAL PAs (at CL2).  However they are not tailored from set of organizational standard processes (CL3).  Let me continue, can you think of a real project plan without processes representing TS, PI, RD?  No, it is nonsense to have a project plan, without requirements analysis or developing a solution (design, coding, etc).  Is there a chance that we can deliver our product to customer without integrating product components?  At ML2, they (processes (tasks) for PAs of ML2 + ML3) all should be there but they are all at CL2.  Spirit of ML3 is not a set of PAs.  What makes an organization at ML3 (spirit) or a process at CL3 (spirit)  is GG3, i.e., standard processes and tailoring guideline.  Management of standard processes is OPD. 

                                OPF should belong ML2.  I will write a CR for that. 
                                In fact, I need to write another CR about making it "explicit" that ML2 = PAs of (ML2+ML3) should be at CL2.  This is now implicit in PP in WBS, Work packages, task descriptions, etc. 

                                ML2 and CL2 is about "living" the processes. 
                                ML3 and CL3 is about "writing" the processes.

                                So, do not start "writing" the processes (tasks) before "living" them.  If you do so, it is highly probable that you will write processes (tasks) unlivable.  And (kudos to Jeff) these processes unfortunately will not set you free, just the opposite.  They will not bring you joy and freedom but misery and slavery.  

                                May the peace and the freedom be upon you,
                                Orhan
                                www.xpi.ca
                                www.nitelik.net


                                2007/5/1, Patrick OToole < PACT.otoole@...>:

                                 
                                Ah, I think I'm finally getting a handle on this discussion.  It seems to me there are two concurrent thought streams running through this dialogue (or at least through my head) - and it all comes down to the phrase "skipping maturity levels."
                                 
                                One interpretation is "let's work on all the ML2 and ML3 process areas at the same time - doing ML2 first, and then working on ML3 afterwards is wimpy - real men do it all!"
                                 
                                The other interpretation is "let's work on SOME of the ML3 things before we work on SOME of the ML2 things.  Afterall, we have project pain that would gain some relief from enhancing our Requirements Elicitation and/or Peer Review processes, and I know that these are ML3 things and that we're an ML1 organization, but still..."
                                 
                                I would STRONGLY agree that typical ML1 organizations are NOT well-served to follow the first interpretation - let's do everything all at once.  Stabilize the project domain before you attack the organizational issues.  Get some credibility and momentum established before you start trying to conquer the world.  (I'll still contend that SOME ML1 organizations, who already have MANY of the ML2 process areas well established are probably in a position where they can start working on some of the ML3 process areas while they are finishing up the ML2 stuff, but that's not what I mean by a "typical ML1 organization.")
                                 
                                Most of my earlier posts were focused on the second interpretation of "skipping maturity levels."  I am certainly NOT a purist when it comes to establishing infrastructure and fixing pain.  If an SEPG-like thing is going to help us get focus and momentum for process improvement, then let's dip into OPF and figure out what an SEPG is and what kinds of things they do.  If OPD can give us a bit of guidance on establishing life cycles, how to write process documentation, and establishing a measurement repository, then let's take a look at that as well.
                                 
                                For pain relief - if improving the peer review process is going to eliminate some pain that we feel in unit, integration, and system testing, then I'm all for it.  If introducing JAD sessions is going to get us better requirements, then I'm all for that too. 
                                 
                                I do not adhere to the philosophy that "you must ONLY do ML2 stuff before you work on any of the ML3 stuff" - and if THAT's what someone means by "skipping maturity levels" then they'd better be prepared for a vigorous debate. 
                                 
                                We must also keep in mind, as I've mentioned before, that organizations may already have a bunch of good stuff in place when they first pick up the model.  They may already have a Project Management Office that has some good risk management practices and may be running IPM-like integrated programs.  They may have an independent test group that is doing their portion of VER extremely well.  They may have a separate training organization that is already outstanding at many of the OT practices.  I would NOT tell these organizations to "quit wasting their time improving these ML3 like practices because there are some gaps at ML2."
                                 
                                One final thought - with 11 process areas, I would contend that ML3 is way too much for a typical ML2 organization to bite off all at once.  The typical ML2 organization might be better served to focus on 3-5 ML3 process areas first - and should probably first focus on those where they perceive the greatest pain reduction or additional benefit.  I'll contend that the typical ML1 organization would be far better served to focus on 2-3 ML2 process areas rather than forming 7 groups, which just HAPPEN to have the same names as each o the ML2 process areas!
                                 
                                I've never been a big believer in trying to achieve maturity levels as quickly as possible.  I would much prefer that the organization focus on gaining the benefits for the projects in a sustainable way than gaining another certificate for the trophy case.  After all, certifications don't produce software - people and projects do.
                                 
                                Regards,
                                 
                                Pat
                                 
                                 
                                 
                                 
                                ----- Original Message -----
                                From: R C Goyal
                                Sent: Monday, April 30, 2007 10:54 PM
                                Subject: Re: {Disarmed} Re: [CMMi Process Improvement] Skipping ML2

                                Dear Orhan

                                Without entering into lengthy discussions may I mention:

                                • Measure about process being counterproductive is not easy. It may also be one's own view point depending upon one's own expectation
                                • Again I would say going from ML1 to ML3  does not mean skipping ML2, however it would also depend the meticulous gap-analysis and improvements for correctly satisfying goals of ML2-PAs and to great extent would depend on  quality of Class C (& B) appraisal format and earnest execution.
                                • In my opinion you can not  satisfy ML3-PA's-goals (how well institutionalized) without ML2 (what)
                                Please comment on proper alignment of my thought process?

                                I always say that a model is means to achieve improvement (even excellence) and the model is not the destination.

                                By  deciding  to go CMMI way , you got some directions for your journey (e.g. L5 which is optimizing a indication of continual process)

                                Regards Goyal


                                From: cmmi_process_improvement@yahoogroups.com [mailto:cmmi_process_improvement@yahoogroups.com] On Behalf Of Orhan KALAYCI
                                Sent: Monday, April 30, 2007 9:59 AM
                                To: cmmi_process_improvement@yahoogroups.com
                                Subject: {Disarmed} Re: [CMMi Process Improvement] Skipping ML2

                                Dear Goyal,

                                This is a good question.  And In fact, this question explains why we are (or I am alone) debating about it.

                                I am talking about this subject over and over because the importance of not skipping levels is widely over looked.  Going from ML1 to ML3 directly is counterproductive.  I believe there are a lot of companies at ML1 jump ML3.  Although these companies certified as ML3, their productivity do not increase even decrease. Then it looks like CMMI is counterproductive.  What is worse, people do not know that this counter productivity is a result of skipping levels but they think CMMI makes these companies less productive. 

                                May the peace be upon you,
                                Orhan
                                www.xpi.ca
                                www.nitelik.net


                                2007/4/30, R C Goyal < rcgoyal@...>:

                                Actually when an organization is being appraised for higher level the appraisal is in any case being done for satisfying goals of all PAs even for lower levels. Then why this subject of skipping is being debated at all.



                                On 4/29/07, Orhan KALAYCI < orhan.kalayci@...> wrote:

                                Hello,

                                I have difficulty in understanding why the most experienced CMMI gurus do not support my proposal:  Let us not encourage organizations to skip ML2! 

                                This is counter productive!  Do you really believe in it?  Do you support not skipping MLs especially ML2?  (We see nobody targeting directly ML5, it is ML2, the most skipped)

                                My questions:
                                -  What value do you think ML2 provides organizations with?
                                -  Do you think skipping ML2 is really counterproductive?  What kind of counter-productiveness does it create?
                                -  How can an organization understand it is ready to start ML3 activities?  Not skipping ML2?

                                May the peace be upon you,
                                Orhan
                                www.xpi.ca
                                www.nitelik.net

                                You know my analogy:  "Before federal and state laws (standard processes) take over, we need to let sheriffs (PMs) to rule their counties (Projects) for a while.  We know, this approach has been proved to be successful in USA.  When ML2 tried to be skipped in Soviet Unions, where they rushed to standardize everything to catch the economics in scale.  We now know this approach (skipping ML2) is not successful, on the opposite it is really counterproductive.  At first sight, it looks like standards should provide productivity, because everybody will work in the same way, synchronization,  harmony, etc will be established.  However, if you rush to standardize too quick.  It results in just the opposite.  As I keep saying, life is full of contradictions."




                                --
                                Thanks & Regards R C Goyal
                                Management Committee  Member,  PMI & CSI Mumbai Chapter
                                BE AN AMBASSADOR OF QUALITY
                                programs@..., rcgoyal@...; rcgoyal@...;
                                Mob: 00919869463964; 00912267108110





                                --
                                Thanks & Regards R C Goyal
                                Management Committee  Member,  PMI & CSI Mumbai Chapter
                                BE AN AMBASSADOR OF QUALITY
                                programs@..., rcgoyal@...; rcgoyal@...;
                                Mob: 00919869463964; 00912267108110





                                --
                                Thanks & Regards R C Goyal
                                Management Committee  Member,  PMI & CSI Mumbai Chapter
                                BE AN AMBASSADOR OF QUALITY
                                programs@..., rcgoyal@...; rcgoyal@...;
                                Mob: 00919869463964; 00912267108110


                              • Orhan KALAYCI
                                Pat, No, I don t think they need to abandon anything already done. But I believe it is better to let people at a ML1 org. know they have right to re-write
                                Message 15 of 30 , May 3, 2007
                                  Pat,

                                  No, I don't think they need to abandon anything already done.  But I believe it is better to let people at a ML1 org. know they have right to re-write these documents according to needs of their projects.  In other words, tailoring in this organization (ML1) should provide projects with really huge space to adapt these docs to their specific needs.

                                  May the peace be upon you,
                                  Orhan
                                  www.xpi.ca
                                  www.nitelik.net


                                  2007/5/3, Patrick OToole < PACT.otoole@...>:

                                   
                                  Orhan,
                                   
                                  If an ML1 organization already has documented processes for its software and test engineering activities, and these documented processes include ways to tailor them to best meet their needs, should they abandon them until they completely implement all practices associated with ML2?
                                   
                                  Regards,
                                   
                                  Pat
                                   
                                   
                                  ----- Original Message -----
                                  Sent: Wednesday, May 02, 2007 1:29 PM
                                  Subject: Re: {Disarmed} Re: [CMMi Process Improvement] Skipping ML2

                                  Hello group, how are you?

                                  To be able discuss or debate on ML2, ML3 or CL2, CL3.  Let me define what they mean:
                                  First of all, an ML is not equal to a set of PAs i.e. ML <> {PAs}

                                  CMMI has a spirit, so do each ML and CL. 

                                  ML2 means "Project management" is in place including processes (tasks) for process areas of ML3 in the project plans. However, they all are at the level of CL2.  We can never think about a project plan without validation or verification tasks in it.  So at ML2, it is for sure there are processes representing VER and VAL PAs (at CL2).  However they are not tailored from set of organizational standard processes (CL3).  Let me continue, can you think of a real project plan without processes representing TS, PI, RD?  No, it is nonsense to have a project plan, without requirements analysis or developing a solution (design, coding, etc).  Is there a chance that we can deliver our product to customer without integrating product components?  At ML2, they (processes (tasks) for PAs of ML2 + ML3) all should be there but they are all at CL2.  Spirit of ML3 is not a set of PAs.  What makes an organization at ML3 (spirit) or a process at CL3 (spirit)  is GG3, i.e., standard processes and tailoring guideline.  Management of standard processes is OPD. 

                                  OPF should belong ML2.  I will write a CR for that. 
                                  In fact, I need to write another CR about making it "explicit" that ML2 = PAs of (ML2+ML3) should be at CL2.  This is now implicit in PP in WBS, Work packages, task descriptions, etc. 

                                  ML2 and CL2 is about "living" the processes. 
                                  ML3 and CL3 is about "writing" the processes.

                                  So, do not start "writing" the processes (tasks) before "living" them.  If you do so, it is highly probable that you will write processes (tasks) unlivable.  And (kudos to Jeff) these processes unfortunately will not set you free, just the opposite.  They will not bring you joy and freedom but misery and slavery.  

                                  May the peace and the freedom be upon you,
                                  Orhan
                                  www.xpi.ca
                                  www.nitelik.net


                                  2007/5/1, Patrick OToole < PACT.otoole@...>:

                                   
                                  Ah, I think I'm finally getting a handle on this discussion.  It seems to me there are two concurrent thought streams running through this dialogue (or at least through my head) - and it all comes down to the phrase "skipping maturity levels."
                                   
                                  One interpretation is "let's work on all the ML2 and ML3 process areas at the same time - doing ML2 first, and then working on ML3 afterwards is wimpy - real men do it all!"
                                   
                                  The other interpretation is "let's work on SOME of the ML3 things before we work on SOME of the ML2 things.  Afterall, we have project pain that would gain some relief from enhancing our Requirements Elicitation and/or Peer Review processes, and I know that these are ML3 things and that we're an ML1 organization, but still..."
                                   
                                  I would STRONGLY agree that typical ML1 organizations are NOT well-served to follow the first interpretation - let's do everything all at once.  Stabilize the project domain before you attack the organizational issues.  Get some credibility and momentum established before you start trying to conquer the world.  (I'll still contend that SOME ML1 organizations, who already have MANY of the ML2 process areas well established are probably in a position where they can start working on some of the ML3 process areas while they are finishing up the ML2 stuff, but that's not what I mean by a "typical ML1 organization.")
                                   
                                  Most of my earlier posts were focused on the second interpretation of "skipping maturity levels."  I am certainly NOT a purist when it comes to establishing infrastructure and fixing pain.  If an SEPG-like thing is going to help us get focus and momentum for process improvement, then let's dip into OPF and figure out what an SEPG is and what kinds of things they do.  If OPD can give us a bit of guidance on establishing life cycles, how to write process documentation, and establishing a measurement repository, then let's take a look at that as well.
                                   
                                  For pain relief - if improving the peer review process is going to eliminate some pain that we feel in unit, integration, and system testing, then I'm all for it.  If introducing JAD sessions is going to get us better requirements, then I'm all for that too. 
                                   
                                  I do not adhere to the philosophy that "you must ONLY do ML2 stuff before you work on any of the ML3 stuff" - and if THAT's what someone means by "skipping maturity levels" then they'd better be prepared for a vigorous debate. 
                                   
                                  We must also keep in mind, as I've mentioned before, that organizations may already have a bunch of good stuff in place when they first pick up the model.  They may already have a Project Management Office that has some good risk management practices and may be running IPM-like integrated programs.  They may have an independent test group that is doing their portion of VER extremely well.  They may have a separate training organization that is already outstanding at many of the OT practices.  I would NOT tell these organizations to "quit wasting their time improving these ML3 like practices because there are some gaps at ML2."
                                   
                                  One final thought - with 11 process areas, I would contend that ML3 is way too much for a typical ML2 organization to bite off all at once.  The typical ML2 organization might be better served to focus on 3-5 ML3 process areas first - and should probably first focus on those where they perceive the greatest pain reduction or additional benefit.  I'll contend that the typical ML1 organization would be far better served to focus on 2-3 ML2 process areas rather than forming 7 groups, which just HAPPEN to have the same names as each o the ML2 process areas!
                                   
                                  I've never been a big believer in trying to achieve maturity levels as quickly as possible.  I would much prefer that the organization focus on gaining the benefits for the projects in a sustainable way than gaining another certificate for the trophy case.  After all, certifications don't produce software - people and projects do.
                                   
                                  Regards,
                                   
                                  Pat
                                   
                                   
                                   
                                   
                                  ----- Original Message -----
                                  From: R C Goyal
                                  Sent: Monday, April 30, 2007 10:54 PM
                                  Subject: Re: {Disarmed} Re: [CMMi Process Improvement] Skipping ML2

                                  Dear Orhan

                                  Without entering into lengthy discussions may I mention:

                                  • Measure about process being counterproductive is not easy. It may also be one's own view point depending upon one's own expectation
                                  • Again I would say going from ML1 to ML3  does not mean skipping ML2, however it would also depend the meticulous gap-analysis and improvements for correctly satisfying goals of ML2-PAs and to great extent would depend on  quality of Class C (& B) appraisal format and earnest execution.
                                  • In my opinion you can not  satisfy ML3-PA's-goals (how well institutionalized) without ML2 (what)
                                  Please comment on proper alignment of my thought process?

                                  I always say that a model is means to achieve improvement (even excellence) and the model is not the destination.

                                  By  deciding  to go CMMI way , you got some directions for your journey (e.g. L5 which is optimizing a indication of continual process)

                                  Regards Goyal


                                  From: cmmi_process_improvement@yahoogroups.com [mailto:cmmi_ process_improvement@yahoogroups.com] On Behalf Of Orhan KALAYCI
                                  Sent: Monday, April 30, 2007 9:59 AM
                                  To: cmmi_process_improvement@yahoogroups.com
                                  Subject: {Disarmed} Re: [CMMi Process Improvement] Skipping ML2

                                  Dear Goyal,

                                  This is a good question.  And In fact, this question explains why we are (or I am alone) debating about it.

                                  I am talking about this subject over and over because the importance of not skipping levels is widely over looked.  Going from ML1 to ML3 directly is counterproductive.  I believe there are a lot of companies at ML1 jump ML3.  Although these companies certified as ML3, their productivity do not increase even decrease. Then it looks like CMMI is counterproductive.  What is worse, people do not know that this counter productivity is a result of skipping levels but they think CMMI makes these companies less productive. 

                                  May the peace be upon you,
                                  Orhan
                                  www.xpi.ca
                                  www.nitelik.net


                                  2007/4/30, R C Goyal < rcgoyal@...>:

                                  Actually when an organization is being appraised for higher level the appraisal is in any case being done for satisfying goals of all PAs even for lower levels. Then why this subject of skipping is being debated at all.



                                  On 4/29/07, Orhan KALAYCI < orhan.kalayci@...> wrote:

                                  Hello,

                                  I have difficulty in understanding why the most experienced CMMI gurus do not support my proposal:  Let us not encourage organizations to skip ML2! 

                                  This is counter productive!  Do you really believe in it?  Do you support not skipping MLs especially ML2?  (We see nobody targeting directly ML5, it is ML2, the most skipped)

                                  My questions:
                                  -  What value do you think ML2 provides organizations with?
                                  -  Do you think skipping ML2 is really counterproductive?  What kind of counter-productiveness does it create?
                                  -  How can an organization understand it is ready to start ML3 activities?  Not skipping ML2?

                                  May the peace be upon you,
                                  Orhan
                                  www.xpi.ca
                                  www.nitelik.net

                                  You know my analogy:  "Before federal and state laws (standard processes) take over, we need to let sheriffs (PMs) to rule their counties (Projects) for a while.  We know, this approach has been proved to be successful in USA.  When ML2 tried to be skipped in Soviet Unions, where they rushed to standardize everything to catch the economics in scale.  We now know this approach (skipping ML2) is not successful, on the opposite it is really counterproductive.  At first sight, it looks like standards should provide productivity, because everybody will work in the same way, synchronization,  harmony, etc will be established.  However, if you rush to standardize too quick.  It results in just the opposite.  As I keep saying, life is full of contradictions."




                                  --
                                  Thanks & Regards R C Goyal
                                  Management Committee  Member,  PMI & CSI Mumbai Chapter
                                  BE AN AMBASSADOR OF QUALITY
                                  programs@... , rcgoyal@...; rcgoyal@...;
                                  Mob: 00919869463964; 00912267108110





                                  --
                                  Thanks & Regards R C Goyal
                                  Management Committee  Member,  PMI & CSI Mumbai Chapter
                                  BE AN AMBASSADOR OF QUALITY
                                  programs@..., rcgoyal@...; rcgoyal@...;
                                  Mob: 00919869463964; 00912267108110



                                • Orhan KALAYCI
                                  Hi group, how are you? Thanks to these posts. I believe it is now clear that there are two dimensions of skipping levels. From ML1 - ML3 or from CL1 - CL3.
                                  Message 16 of 30 , May 3, 2007
                                    Hi group, how are you?

                                    Thanks to these posts. I believe it is now clear that there are two dimensions of skipping levels. 
                                    From ML1 -> ML3 or from CL1 -> CL3.

                                    When we say ML1->ML3 in fact this is a short way of saying a set of PAs going from CL1->CL3. 
                                    If fact, when you zoom into this situation real problem is going from CL1->CL3.  So if an PA is already at CL2, of course there is no problem for this specific PA to continue towards CL3.  However, be aware of the problem of synchronization among other PAs that may at CL2. 

                                    Let us not forget the spirit:
                                    Let us not encourage people writing processes in fact when they do not know these processes. In other words, when they are not ready to write the standard processes for others to follow.  First, they should discover the processes by not reading the books about them but living in these processes for a while (ML2).  At ML2, we just name the processes and ask people to develop their own understanding by try it live it and learn it.  

                                    May the peace be upon you,
                                    Orhan
                                    www.xpi.ca
                                    www.nitelik.net

                                    2007/5/3, Patrick OToole < PACT.otoole@...>:

                                     
                                    Everyone,
                                     
                                    Here's another "problem" that I'm having with the "don't do things at ML3 until you've got ML2 nailed" position.  Let's say an organization has staffed its process improvement project such that it can focus on improving 7 things concurrently.  Let's say they initially form one team to work on REQM, another to work on PP, another to work on PMC, etc.
                                     
                                    But, in this organization, project management is already in pretty good shape, while MA and PPQA are pretty much starting from scratch.  In such a case, the PP and PMC teams may wrap up pretty quickly - within a month or two they may fill the couple of gaps, train the other PMs to adopt the new practices in a value-added way and they're done.
                                     
                                    In the meantime, the MA and PPQA teams are trying to understand what the associated practices are all about, they are off getting trained on their respective process areas, and have put together a one-year plan to become "model compliant."
                                     
                                    Since the organization is already staffed to run 7 concurrent improvement teams, they COULD re-allocate all the PP and PMC folks to assist with MA and PPQA, and they'd wind up with twice as many people in a group grope.  The PP and PMC teams would now need to be trained and would be playing 2 months of catch up.  This MAY reduce the 12 month schedule by a month or two - or it may not.
                                     
                                    Alternatively, the PP and PMC teams could be re-directed to work on RSKM and IPM, maybe targeting CL1 or CL2 in these two process areas.  After all, this is simply an extension of the project management work they have already performed AND there may be a pretty solid foundation in some of these practices as well.
                                     
                                    Would the organization be WRONG if they decided to go with the second option?
                                     
                                    And, if they get IPM and RSKM to CL2 within 6 months, would they be wrong to then start working on PP and PMC GP3.1 and GP3.2?  Is there anything wrong with trying to standardize the practices and processes that were pretty much deeply institutionalized anyway?
                                     
                                    As much as I agree with the basic principle that we "shouldn't skip maturity levels," I'm merely suggesting that it is hardly ever that simple, and that we really need to look at each unique situation to determine what's best for THAT organization.  Just as there is no one way to implement any given CMMI practice, and we encourage organizations to implement the practices in a way that best fits their needs, shouldn't we be somewhat flexible in our statements about the ORDER in which they implement the practices?
                                     
                                    And on it goes...
                                     
                                    Regards,
                                     
                                    Pat
                                     
                                     
                                     
                                     
                                    ----- Original Message -----
                                    Sent: Wednesday, May 02, 2007 1:29 PM
                                    Subject: Re: {Disarmed} Re: [CMMi Process Improvement] Skipping ML2

                                    Hello group, how are you?

                                    To be able discuss or debate on ML2, ML3 or CL2, CL3.  Let me define what they mean:
                                    First of all, an ML is not equal to a set of PAs i.e. ML <> {PAs}

                                    CMMI has a spirit, so do each ML and CL. 

                                    ML2 means "Project management" is in place including processes (tasks) for process areas of ML3 in the project plans. However, they all are at the level of CL2.  We can never think about a project plan without validation or verification tasks in it.  So at ML2, it is for sure there are processes representing VER and VAL PAs (at CL2).  However they are not tailored from set of organizational standard processes (CL3).  Let me continue, can you think of a real project plan without processes representing TS, PI, RD?  No, it is nonsense to have a project plan, without requirements analysis or developing a solution (design, coding, etc).  Is there a chance that we can deliver our product to customer without integrating product components?  At ML2, they (processes (tasks) for PAs of ML2 + ML3) all should be there but they are all at CL2.  Spirit of ML3 is not a set of PAs.  What makes an organization at ML3 (spirit) or a process at CL3 (spirit)  is GG3, i.e., standard processes and tailoring guideline.  Management of standard processes is OPD. 

                                    OPF should belong ML2.  I will write a CR for that. 
                                    In fact, I need to write another CR about making it "explicit" that ML2 = PAs of (ML2+ML3) should be at CL2.  This is now implicit in PP in WBS, Work packages, task descriptions, etc. 

                                    ML2 and CL2 is about "living" the processes. 
                                    ML3 and CL3 is about "writing" the processes.

                                    So, do not start "writing" the processes (tasks) before "living" them.  If you do so, it is highly probable that you will write processes (tasks) unlivable.  And (kudos to Jeff) these processes unfortunately will not set you free, just the opposite.  They will not bring you joy and freedom but misery and slavery.  

                                    May the peace and the freedom be upon you,
                                    Orhan
                                    www.xpi.ca
                                    www.nitelik.net


                                    2007/5/1, Patrick OToole < PACT.otoole@...>:

                                     
                                    Ah, I think I'm finally getting a handle on this discussion.  It seems to me there are two concurrent thought streams running through this dialogue (or at least through my head) - and it all comes down to the phrase "skipping maturity levels."
                                     
                                    One interpretation is "let's work on all the ML2 and ML3 process areas at the same time - doing ML2 first, and then working on ML3 afterwards is wimpy - real men do it all!"
                                     
                                    The other interpretation is "let's work on SOME of the ML3 things before we work on SOME of the ML2 things.  Afterall, we have project pain that would gain some relief from enhancing our Requirements Elicitation and/or Peer Review processes, and I know that these are ML3 things and that we're an ML1 organization, but still..."
                                     
                                    I would STRONGLY agree that typical ML1 organizations are NOT well-served to follow the first interpretation - let's do everything all at once.  Stabilize the project domain before you attack the organizational issues.  Get some credibility and momentum established before you start trying to conquer the world.  (I'll still contend that SOME ML1 organizations, who already have MANY of the ML2 process areas well established are probably in a position where they can start working on some of the ML3 process areas while they are finishing up the ML2 stuff, but that's not what I mean by a "typical ML1 organization.")
                                     
                                    Most of my earlier posts were focused on the second interpretation of "skipping maturity levels."  I am certainly NOT a purist when it comes to establishing infrastructure and fixing pain.  If an SEPG-like thing is going to help us get focus and momentum for process improvement, then let's dip into OPF and figure out what an SEPG is and what kinds of things they do.  If OPD can give us a bit of guidance on establishing life cycles, how to write process documentation, and establishing a measurement repository, then let's take a look at that as well.
                                     
                                    For pain relief - if improving the peer review process is going to eliminate some pain that we feel in unit, integration, and system testing, then I'm all for it.  If introducing JAD sessions is going to get us better requirements, then I'm all for that too. 
                                     
                                    I do not adhere to the philosophy that "you must ONLY do ML2 stuff before you work on any of the ML3 stuff" - and if THAT's what someone means by "skipping maturity levels" then they'd better be prepared for a vigorous debate. 
                                     
                                    We must also keep in mind, as I've mentioned before, that organizations may already have a bunch of good stuff in place when they first pick up the model.  They may already have a Project Management Office that has some good risk management practices and may be running IPM-like integrated programs.  They may have an independent test group that is doing their portion of VER extremely well.  They may have a separate training organization that is already outstanding at many of the OT practices.  I would NOT tell these organizations to "quit wasting their time improving these ML3 like practices because there are some gaps at ML2."
                                     
                                    One final thought - with 11 process areas, I would contend that ML3 is way too much for a typical ML2 organization to bite off all at once.  The typical ML2 organization might be better served to focus on 3-5 ML3 process areas first - and should probably first focus on those where they perceive the greatest pain reduction or additional benefit.  I'll contend that the typical ML1 organization would be far better served to focus on 2-3 ML2 process areas rather than forming 7 groups, which just HAPPEN to have the same names as each o the ML2 process areas!
                                     
                                    I've never been a big believer in trying to achieve maturity levels as quickly as possible.  I would much prefer that the organization focus on gaining the benefits for the projects in a sustainable way than gaining another certificate for the trophy case.  After all, certifications don't produce software - people and projects do.
                                     
                                    Regards,
                                     
                                    Pat
                                     
                                     
                                     
                                     
                                    ----- Original Message -----
                                    From: R C Goyal
                                    Sent: Monday, April 30, 2007 10:54 PM
                                    Subject: Re: {Disarmed} Re: [CMMi Process Improvement] Skipping ML2

                                    Dear Orhan

                                    Without entering into lengthy discussions may I mention:

                                    • Measure about process being counterproductive is not easy. It may also be one's own view point depending upon one's own expectation
                                    • Again I would say going from ML1 to ML3  does not mean skipping ML2, however it would also depend the meticulous gap-analysis and improvements for correctly satisfying goals of ML2-PAs and to great extent would depend on  quality of Class C (& B) appraisal format and earnest execution.
                                    • In my opinion you can not  satisfy ML3-PA's-goals (how well institutionalized) without ML2 (what)
                                    Please comment on proper alignment of my thought process?

                                    I always say that a model is means to achieve improvement (even excellence) and the model is not the destination.

                                    By  deciding  to go CMMI way , you got some directions for your journey (e.g. L5 which is optimizing a indication of continual process)

                                    Regards Goyal


                                    From: cmmi_process_improvement@yahoogroups.com [mailto:cmmi_ process_improvement@yahoogroups.com] On Behalf Of Orhan KALAYCI
                                    Sent: Monday, April 30, 2007 9:59 AM
                                    To: cmmi_process_improvement@yahoogroups.com
                                    Subject: {Disarmed} Re: [CMMi Process Improvement] Skipping ML2

                                    Dear Goyal,

                                    This is a good question.  And In fact, this question explains why we are (or I am alone) debating about it.

                                    I am talking about this subject over and over because the importance of not skipping levels is widely over looked.  Going from ML1 to ML3 directly is counterproductive.  I believe there are a lot of companies at ML1 jump ML3.  Although these companies certified as ML3, their productivity do not increase even decrease. Then it looks like CMMI is counterproductive.  What is worse, people do not know that this counter productivity is a result of skipping levels but they think CMMI makes these companies less productive. 

                                    May the peace be upon you,
                                    Orhan
                                    www.xpi.ca
                                    www.nitelik.net


                                    2007/4/30, R C Goyal < rcgoyal@...>:

                                    Actually when an organization is being appraised for higher level the appraisal is in any case being done for satisfying goals of all PAs even for lower levels. Then why this subject of skipping is being debated at all.



                                    On 4/29/07, Orhan KALAYCI < orhan.kalayci@...> wrote:

                                    Hello,

                                    I have difficulty in understanding why the most experienced CMMI gurus do not support my proposal:  Let us not encourage organizations to skip ML2! 

                                    This is counter productive!  Do you really believe in it?  Do you support not skipping MLs especially ML2?  (We see nobody targeting directly ML5, it is ML2, the most skipped)

                                    My questions:
                                    -  What value do you think ML2 provides organizations with?
                                    -  Do you think skipping ML2 is really counterproductive?  What kind of counter-productiveness does it create?
                                    -  How can an organization understand it is ready to start ML3 activities?  Not skipping ML2?

                                    May the peace be upon you,
                                    Orhan
                                    www.xpi.ca
                                    www.nitelik.net

                                    You know my analogy:  "Before federal and state laws (standard processes) take over, we need to let sheriffs (PMs) to rule their counties (Projects) for a while.  We know, this approach has been proved to be successful in USA.  When ML2 tried to be skipped in Soviet Unions, where they rushed to standardize everything to catch the economics in scale.  We now know this approach (skipping ML2) is not successful, on the opposite it is really counterproductive.  At first sight, it looks like standards should provide productivity, because everybody will work in the same way, synchronization,  harmony, etc will be established.  However, if you rush to standardize too quick.  It results in just the opposite.  As I keep saying, life is full of contradictions."




                                    --
                                    Thanks & Regards R C Goyal
                                    Management Committee  Member,  PMI & CSI Mumbai Chapter
                                    BE AN AMBASSADOR OF QUALITY
                                    programs@... , rcgoyal@...; rcgoyal@...;
                                    Mob: 00919869463964; 00912267108110





                                    --
                                    Thanks & Regards R C Goyal
                                    Management Committee  Member,  PMI & CSI Mumbai Chapter
                                    BE AN AMBASSADOR OF QUALITY
                                    programs@..., rcgoyal@...; rcgoyal@...;
                                    Mob: 00919869463964; 00912267108110



                                  • Heather Oppenheimer
                                    Here s a summary of some of the key points in this (semi-religious) discussion. At the core, I don t really think people are really disagreeing, although
                                    Message 17 of 30 , May 3, 2007
                                      Here's a summary of some of the key points in this (semi-religious) discussion.  At the core, I don't really think people are really disagreeing, although apparently there  are strong opinions about the way things are said.  The phrases "maturity levels"  and "skipping maturity levels" seem to be the major part of the confusion.  Sorry it's so long, but it's my last post on this topic.
                                       
                                      Agreements so far (these are not necessarily in priority order)
                                      1.  It's important to focus process improvement on areas of business need - usually to address pain. 
                                      2.  It's important to get the management of projects under control.  (In the CMMI model, this usually maps to the specific practices found in the ML2 PAs and/or the generic practices found under GG2 for any/all PAs.)
                                      3.  It's difficult to do a good job on many things at once - the heuristic is that people can handle 5 +- 2 "chunks" at a time. 
                                      4.  It's important to have processes that reflect reality and experience, not processes that are written based on some idea of The Perfect Process. 
                                      5.  It's important to share experience/knowledge/lessons learned among projects.
                                      6. The work done in a project includes more than just the activities related to the PAs grouped at ML2
                                       
                                      Terminology and interpretation disagreements- Maturity Level:
                                      1. The CMMI model says "A maturity level consists of related specific and generic practice for a predefined set of process areas..."
                                      (Interestingly enough though, when it describes the maturity levels, maturity level 2 sounds a lot more like a description of the GG2 generic practices, and only refers to some of the specific practices of the ML2 PAs.)
                                      2. Orhan says "ML2  <> the 7 PAs... the spirit of ML2 is that all processes related to ML2 and ML3 are at CL2". 
                                      3. I think most of the rest of us have been using the definition "Maturity Level 2 = the specific and generic practices of PP, PMC, REQM, CM, PPQA, MA and SAM". 
                                       
                                      Can we agree to use the model definition that ML2 is the specific practices and GG2 generic practices of PP, PMC, REQM, SAM, MA, PPQA, and CM?  And that ML3 is the specific practices and GG2 + GG3 generic practices of the ML2 and ML3 process areas?  Also that CL2 includes the specific practices and GG2 generic practices of whichever PA you are talking about, while CL3 adds the GG3 generic practices?
                                       
                                      Terminology and interpretation disagreements- Skipping Maturity Levels:
                                      1. With respect to skipping maturity levels, the CMMI model says "As your organization achieves the generic and specific goals for the set of process areas in a maturity level, you are increasing your organizational maturity level... Because each maturity level forms a necessary foundation for the next level, trying to skip maturity levels is usually counterproductive."
                                      2. To Orhan,  "skipping maturity levels" means "writing the processes before living them" and "having standard processes before projects have spent time managing their own processes."
                                      3. Pat says about skipping maturity levels: One interpretation is "let's work on all the ML2 and ML3 process areas at the same time -doing ML2 first, and then working on ML3 afterwards is wimpy - real men do it all!"  The other interpretation is "let's work on SOME of the ML3 things before we work on SOME of the ML2 things.  Afterall, we have project pain that would gain some relief from enhancing our Requirements Elicitation and/or Peer Review processes, and I know that these are ML3 things and that we're an ML1 organization, but still..."
                                      4. I have been using the same interpretations as Pat.  I think Orhan's definition is more related to the phrase "skipping capability levels".
                                       
                                      Can we agree that we don't think working on all the ML2 and ML3 process areas at the same time is a good idea?
                                      Do we agree that we think that it's best to focus on making sure processes are managed (CL2) before trying to standardize them (GP3.1)?  What about starting to collect information from projects (GP3.2) even if all of the GG2 practices aren't completely implemented yet?
                                      We haven't talked about CL1... Do we agree that it's best to focus first on performing processes (CL1) before trying to manage them (CL2)???  That would seem to follow the same logic.
                                       
                                      Issues -
                                      I think the only real point of disagreement is whether or not an organization MUST first manage processes (GG2) at the project level and MUST NOT do anything organizationally for those processes until they are managed in each project. 
                                       
                                      1. In some organizations, the best way to ensure both buy in and development of standard processes that can be appropriately tailored, is to allow each project the freedom to implement their own processes first, then synthesize or select the "best".
                                      2. In some organizations, people work on more than one project and they need process consistency across those projects so they can be effective.
                                      3. In some organizations, people are successfully using some standard processes before they start working with CMMI (e.g processes related to configuration management, processes related to managing suppliers...). It doesn't make sense to stop using those processes and create new project-specific ones.
                                      4. In some organizations, the skills and expertise needed for developing processes (knowledge mining, domain expertise, communication, leadership...) are not available in every project.
                                       
                                      The CMMI model says the process descriptions and procedures may be quite different in each specific instance at CL2/ML2 - I think it means it's ok if they're different, it's ok if they are based on an organizational standard.
                                       
                                      Can we agree that that  "it depends" is the answer to "When and how should an organization implement standard processes?"
                                       
                                      Heather Oppenheimer
                                       
                                       
                                       
                                       
                                    • Andre Heijstek
                                      Hi all, Some remarks. First of all, I like your notion of the spirit of a maturity level. I think you are absolutely right in what you are stating here.
                                      Message 18 of 30 , May 3, 2007
                                        Re: {Disarmed} Re: [CMMi Process Improvement] Skipping ML2 Hi all,

                                        Some remarks.

                                        First of all, I like your notion of the spirit of a maturity level. I think you are absolutely right in what you are stating here.

                                        Secondly, it always helps me to think of CMMI as an improvement model, not as a process model. In an improvement model the maturity levels tell me in what order I have to IMPROVE things, not in what order I have to DO things.
                                        So, of course you will DO a lot of TS and VER being a ML1 organization, and of course you will have to continue doing that when working towards ML2. You will also DO quite a bit of OPF when working from ML1 to ML2. But you probably won’t get to master everything OPF expects in your ML2 project. So mastering OPF is left as an exercise for your ML3 project.

                                        Your statements about living and writing the processes are not completely correct in my view.
                                        In ML2 you have to write processes anyway before you can systematically ‘live’ them. You could however document them at the project level (as part of your project plan for example), so they don’t have to be organizational processes.
                                        In ML3 you are writing org. wide processes, but have to live these as well.

                                        Regards, Andre.

                                        On 02-05-2007 20:29, "Orhan KALAYCI" <orhan.kalayci@...> wrote:


                                         
                                         

                                        Hello group, how are you?

                                        To be able discuss or debate on ML2, ML3 or CL2, CL3.  Let me define what they mean:
                                        First of all, an ML is not equal to a set of PAs i.e. ML <> {PAs}

                                        CMMI has a spirit, so do each ML and CL.  

                                        ML2 means "Project management" is in place including processes (tasks) for process areas of ML3 in the project plans. However, they all are at the level of CL2.  We can never think about a project plan without validation or verification tasks in it.  So at ML2, it is for sure there are processes representing VER and VAL PAs (at CL2).  However they are not tailored from set of organizational standard processes (CL3).  Let me continue, can you think of a real project plan without processes representing TS, PI, RD?  No, it is nonsense to have a project plan, without requirements analysis or developing a solution (design, coding, etc).  Is there a chance that we can deliver our product to customer without integrating product components?  At ML2, they (processes (tasks) for PAs of ML2 + ML3) all should be there but they are all at CL2.  Spirit of ML3 is not a set of PAs.  What makes an organization at ML3 (spirit) or a process at CL3 (spirit)  is GG3, i.e., standard processes and tailoring guideline.  Management of standard processes is OPD.  

                                        OPF should belong ML2.  I will write a CR for that.  
                                        In fact, I need to write another CR about making it "explicit" that ML2 = PAs of (ML2+ML3) should be at CL2.  This is now implicit in PP in WBS, Work packages, task descriptions, etc.  

                                        ML2 and CL2 is about "living" the processes.  
                                        ML3 and CL3 is about "writing" the processes.

                                        So, do not start "writing" the processes (tasks) before "living" them.  If you do so, it is highly probable that you will write processes (tasks) unlivable.  And (kudos to Jeff) these processes unfortunately will not set you free, just the opposite.  They will not bring you joy and freedom but misery and slavery.    

                                        May the peace and the freedom be upon you,
                                        Orhan
                                        www.xpi.ca <http://www.xpi.ca>
                                        www.nitelik.net <http://www.nitelik.net>


                                        2007/5/1, Patrick OToole < PACT.otoole@...>:

                                         
                                         

                                         
                                        Ah, I think I'm finally getting a handle on this discussion.  It seems to me there are two concurrent thought streams running through this dialogue (or at least through my head) - and it all comes down to the phrase "skipping maturity levels."

                                        One interpretation is "let's work on all the ML2 and ML3 process areas at the same time - doing ML2 first, and then working on ML3 afterwards is wimpy - real men do it all!"

                                        The other interpretation is "let's work on SOME of the ML3 things before we work on SOME of the ML2 things.  Afterall, we have project pain that would gain some relief from enhancing our Requirements Elicitation and/or Peer Review processes, and I know that these are ML3 things and that we're an ML1 organization, but still..."

                                        I would STRONGLY agree that typical ML1 organizations are NOT well-served to follow the first interpretation - let's do everything all at once.  Stabilize the project domain before you attack the organizational issues.  Get some credibility and momentum established before you start trying to conquer the world.  (I'll still contend that SOME ML1 organizations, who already have MANY of the ML2 process areas well established are probably in a position where they can start working on some of the ML3 process areas while they are finishing up the ML2 stuff, but that's not what I mean by a "typical ML1 organization.")

                                        Most of my earlier posts were focused on the second interpretation of "skipping maturity levels."  I am certainly NOT a purist when it comes to establishing infrastructure and fixing pain.  If an SEPG-like thing is going to help us get focus and momentum for process improvement, then let's dip into OPF and figure out what an SEPG is and what kinds of things they do.  If OPD can give us a bit of guidance on establishing life cycles, how to write process documentation, and establishing a measurement repository, then let's take a look at that as well.

                                        For pain relief - if improving the peer review process is going to eliminate some pain that we feel in unit, integration, and system testing, then I'm all for it.  If introducing JAD sessions is going to get us better requirements, then I'm all for that too.  

                                        I do not adhere to the philosophy that "you must ONLY do ML2 stuff before you work on any of the ML3 stuff" - and if THAT's what someone means by "skipping maturity levels" then they'd better be prepared for a vigorous debate.
                                         
                                        We must also keep in mind, as I've mentioned before, that organizations may already have a bunch of good stuff in place when they first pick up the model.  They may already have a Project Management Office that has some good risk management practices and may be running IPM-like integrated programs.  They may have an independent test group that is doing their portion of VER extremely well.  They may have a separate training organization that is already outstanding at many of the OT practices.  I would NOT tell these organizations to "quit wasting their time improving these ML3 like practices because there are some gaps at ML2."

                                        One final thought - with 11 process areas, I would contend that ML3 is way too much for a typical ML2 organization to bite off all at once.  The typical ML2 organization might be better served to focus on 3-5 ML3 process areas first - and should probably first focus on those where they perceive the greatest pain reduction or additional benefit.  I'll contend that the typical ML1 organization would be far better served to focus on 2-3 ML2 process areas rather than forming 7 groups, which just HAPPEN to have the same names as each o the ML2 process areas!

                                        I've never been a big believer in trying to achieve maturity levels as quickly as possible.  I would much prefer that the organization focus on gaining the benefits for the projects in a sustainable way than gaining another certificate for the trophy case.  After all, certifications don't produce software - people and projects do.

                                        Regards,

                                        Pat

                                         
                                         
                                         

                                        ----- Original Message -----
                                         
                                        From:  R C Goyal <mailto:rcgoyal@...>  
                                         
                                        To: cmmi_process_improvement@yahoogroups.com  
                                         
                                        Sent: Monday, April 30, 2007 10:54  PM
                                         
                                        Subject: Re: {Disarmed} Re: [CMMi Process  Improvement] Skipping ML2
                                         

                                         
                                         

                                        Dear Orhan

                                        Without entering into lengthy discussions may I  mention:
                                         
                                        • Measure about process being counterproductive is not easy. It may also  be one's own view point depending upon one's own expectation  
                                        • Again I would say going from ML1 to ML3  does not mean skipping  ML2, however it would also depend the meticulous gap-analysis and  improvements for correctly satisfying goals of ML2-PAs and to great extent  would depend on  quality of Class C (& B) appraisal format and  earnest execution.  
                                        • In my opinion you can not  satisfy ML3-PA's-goals (how well  institutionalized) without ML2 (what)
                                        Please comment on proper  alignment of my thought process?

                                        I always say that a model is means to  achieve improvement (even excellence) and the model is not the destination.  

                                        By  deciding  to go CMMI way , you got some directions for  your journey (e.g. L5 which is optimizing a indication of continual  process)

                                        Regards Goyal
                                         
                                         

                                         
                                         

                                         

                                         

                                        From: cmmi_process_improvement@yahoogroups.com  [mailto:cmmi_ process_improvement@yahoogroups.com <mailto:process_improvement@yahoogroups.com> ]  On Behalf Of Orhan KALAYCI
                                        Sent: Monday, April 30, 2007  9:59 AM
                                        To: cmmi_process_improvement@yahoogroups.com
                                        Subject:  {Disarmed} Re: [CMMi Process Improvement] Skipping ML2

                                         
                                         
                                         

                                        Dear Goyal,

                                        This is a good question.  And In fact, this  question explains why we are (or I am alone) debating about it.

                                        I  am talking about this subject over and over because the importance of not  skipping levels is widely over looked.  
                                        Going from  ML1 to ML3 directly is counterproductive. I believe there are  a lot of companies at ML1 jump ML3.  Although these companies  certified as ML3, their productivity do not increase even decrease. Then it looks  like CMMI is counterproductive. What is worse, people do not  know that this counter productivity is a result of  skipping levels but they think CMMI makes these companies  less productive.  

                                        May the peace be upon you,
                                        Orhan
                                        www.xpi.ca <http://www.xpi.ca>
                                        www.nitelik.net <http://www.nitelik.net>


                                         
                                        2007/4/30, R C Goyal <  rcgoyal@... <mailto:rcgoyal@...> >:  

                                         
                                         
                                         
                                         

                                        Actually when an organization is being appraised for  higher level the appraisal is in any case being done for satisfying  goals of all PAs even for lower levels. Then why this subject of skipping is  being debated at all.



                                         
                                        On 4/29/07, Orhan  KALAYCI <  orhan.kalayci@... <mailto:orhan.kalayci@...> > wrote:  

                                         
                                         
                                         
                                         

                                        Hello,

                                        I have difficulty in understanding why the most  experienced CMMI gurus do not support my proposal:  Let us not  encourage organizations to skip ML2!  

                                        This is counter  productive!  Do you really believe in it?  Do you support  not skipping MLs especially ML2?  (We see nobody targeting  directly ML5, it is ML2, the most skipped)

                                        My  questions:
                                        -  What value do you think ML2 provides  organizations with?
                                        -  Do you think skipping ML2 is really  counterproductive?  What kind of counter-productiveness does  it create?
                                        -  How can an organization understand it is ready  to start ML3 activities?  Not skipping ML2?

                                        May the peace  be upon you,
                                        Orhan
                                        www.xpi.ca <http://www.xpi.ca>
                                        www.nitelik.net <http://www.nitelik.net>

                                        You know my  analogy:  "Before federal and state laws (standard processes)  take over, we need to let sheriffs (PMs) to rule their counties  (Projects) for a while.  We know, this approach has been proved  to be successful in USA.  When ML2 tried to be skipped in Soviet  Unions, where they rushed to standardize everything to catch the  economics in scale.  We now know this approach (skipping ML2) is  not successful, on the opposite it is really counterproductive.   At first sight, it looks like standards should provide productivity,  because everybody will work in the same way, synchronization,   harmony, etc will be established.  However, if you rush to  standardize too quick.  It results in just the opposite.  As  I keep saying, life is full of contradictions."



                                      • Orhan KALAYCI
                                        Hi Andre, how are you? Thank you for saying aloud that I am right :) But that is not enough I should be right for everything I said :)) Just for laugh :)
                                        Message 19 of 30 , May 3, 2007
                                          Hi Andre, how are you?

                                          Thank you for saying aloud that I am right :)

                                          But that is not enough I should be right for everything I said :)) 

                                          Just for laugh :)

                                          Writing is an interesting term.  I am writing a book now. It took almost three times longer than I originally estimated or guesstimated :)   I do not know "have you ever write a book?" (May be you have a PhD and you wrote a book then.  If not, you would most probably under estimate how hard it is!  I did! :)  But I like it very much I am planning to write several books in the near future.  This one is in Turkish.  Others will be in English.  I think of a book named "The Spirit of CMMI"

                                          We live in information age.  We never work on or in the ground (as in agricultural ages) or we even do not in a factory floor (as in industrial age).  There are computers in front of all of us.  We keep writing all the day.  Otherwise, we either speak or listen but no digging or using no hard machine, etc... 

                                          So, when I say "writing a process" I mean something similar to "writing a book"  It is extremely difficult because we write it for others to read and learn and hopefully use in their works.  If we do write for ourselves or some small group of colleagues within a single project it is no problem even if there is mistake or something missing we can handle it within the group.  If we write a process and all organization is supposed to be bound with it.  Then it requires a living experience of the topic that we are writing about. 

                                          OPF:  The spirit of OPF is setting targets for process improvement and managing these targets.  Setting process improvement targets begins with ML2.  So it should be placed in ML2.  SG3 of OPF, I think, can be moved to OPD.  OPD now is in a static manner, moving SG3 of OPF to OPD makes it more dynamic.  Do you know, what I realized, I think OPD and OPF are like processes.  This is another spiritual issue:  distinguishing between concepts of "process area" and "process"

                                          May the peace and the scripts and the processes and the process areas be upon you,
                                          Orhan
                                          www.xpi.ca
                                          www.nitelik.net


                                          2007/5/3, Andre Heijstek <andre.heijstek@...>:

                                          Hi all,

                                          Some remarks.

                                          First of all, I like your notion of the spirit of a maturity level. I think you are absolutely right in what you are stating here.

                                          Secondly, it always helps me to think of CMMI as an improvement model, not as a process model. In an improvement model the maturity levels tell me in what order I have to IMPROVE things, not in what order I have to DO things.
                                          So, of course you will DO a lot of TS and VER being a ML1 organization, and of course you will have to continue doing that when working towards ML2. You will also DO quite a bit of OPF when working from ML1 to ML2. But you probably won't get to master everything OPF expects in your ML2 project. So mastering OPF is left as an exercise for your ML3 project.

                                          Your statements about living and writing the processes are not completely correct in my view.
                                          In ML2 you have to write processes anyway before you can systematically 'live' them. You could however document them at the project level (as part of your project plan for example), so they don't have to be organizational processes.
                                          In ML3 you are writing org. wide processes, but have to live these as well.

                                          Regards, Andre.

                                          On 02-05-2007 20:29, "Orhan KALAYCI" <orhan.kalayci@xpi.ca> wrote:


                                           
                                           

                                          Hello group, how are you?

                                          To be able discuss or debate on ML2, ML3 or CL2, CL3.  Let me define what they mean:
                                          First of all, an ML is not equal to a set of PAs i.e. ML <> {PAs}

                                          CMMI has a spirit, so do each ML and CL.  

                                          ML2 means "Project management" is in place including processes (tasks) for process areas of ML3 in the project plans. However, they all are at the level of CL2.  We can never think about a project plan without validation or verification tasks in it.  So at ML2, it is for sure there are processes representing VER and VAL PAs (at CL2).  However they are not tailored from set of organizational standard processes (CL3).  Let me continue, can you think of a real project plan without processes representing TS, PI, RD?  No, it is nonsense to have a project plan, without requirements analysis or developing a solution (design, coding, etc).  Is there a chance that we can deliver our product to customer without integrating product components?  At ML2, they (processes (tasks) for PAs of ML2 + ML3) all should be there but they are all at CL2.  Spirit of ML3 is not a set of PAs.  What makes an organization at ML3 (spirit) or a process at CL3 (spirit)  is GG3, i.e., standard processes and tailoring guideline.  Management of standard processes is OPD.  

                                          OPF should belong ML2.  I will write a CR for that.  
                                          In fact, I need to write another CR about making it "explicit" that ML2 = PAs of (ML2+ML3) should be at CL2.  This is now implicit in PP in WBS, Work packages, task descriptions, etc.  

                                          ML2 and CL2 is about "living" the processes.  
                                          ML3 and CL3 is about "writing" the processes.

                                          So, do not start "writing" the processes (tasks) before "living" them.  If you do so, it is highly probable that you will write processes (tasks) unlivable.  And (kudos to Jeff) these processes unfortunately will not set you free, just the opposite.  They will not bring you joy and freedom but misery and slavery.    

                                          May the peace and the freedom be upon you,
                                          Orhan
                                          www.xpi.ca <http://www.xpi.ca>
                                          www.nitelik.net <http://www.nitelik.net>


                                          2007/5/1, Patrick OToole < PACT.otoole@att.net>:

                                           
                                           

                                           
                                          Ah, I think I'm finally getting a handle on this discussion.  It seems to me there are two concurrent thought streams running through this dialogue (or at least through my head) - and it all comes down to the phrase "skipping maturity levels."

                                          One interpretation is "let's work on all the ML2 and ML3 process areas at the same time - doing ML2 first, and then working on ML3 afterwards is wimpy - real men do it all!"

                                          The other interpretation is "let's work on SOME of the ML3 things before we work on SOME of the ML2 things.  Afterall, we have project pain that would gain some relief from enhancing our Requirements Elicitation and/or Peer Review processes, and I know that these are ML3 things and that we're an ML1 organization, but still..."

                                          I would STRONGLY agree that typical ML1 organizations are NOT well-served to follow the first interpretation - let's do everything all at once.  Stabilize the project domain before you attack the organizational issues.  Get some credibility and momentum established before you start trying to conquer the world.  (I'll still contend that SOME ML1 organizations, who already have MANY of the ML2 process areas well established are probably in a position where they can start working on some of the ML3 process areas while they are finishing up the ML2 stuff, but that's not what I mean by a "typical ML1 organization.")

                                          Most of my earlier posts were focused on the second interpretation of "skipping maturity levels."  I am certainly NOT a purist when it comes to establishing infrastructure and fixing pain.  If an SEPG-like thing is going to help us get focus and momentum for process improvement, then let's dip into OPF and figure out what an SEPG is and what kinds of things they do.  If OPD can give us a bit of guidance on establishing life cycles, how to write process documentation, and establishing a measurement repository, then let's take a look at that as well.

                                          For pain relief - if improving the peer review process is going to eliminate some pain that we feel in unit, integration, and system testing, then I'm all for it.  If introducing JAD sessions is going to get us better requirements, then I'm all for that too.  

                                          I do not adhere to the philosophy that "you must ONLY do ML2 stuff before you work on any of the ML3 stuff" - and if THAT's what someone means by "skipping maturity levels" then they'd better be prepared for a vigorous debate.
                                           
                                          We must also keep in mind, as I've mentioned before, that organizations may already have a bunch of good stuff in place when they first pick up the model.  They may already have a Project Management Office that has some good risk management practices and may be running IPM-like integrated programs.  They may have an independent test group that is doing their portion of VER extremely well.  They may have a separate training organization that is already outstanding at many of the OT practices.  I would NOT tell these organizations to "quit wasting their time improving these ML3 like practices because there are some gaps at ML2."

                                          One final thought - with 11 process areas, I would contend that ML3 is way too much for a typical ML2 organization to bite off all at once.  The typical ML2 organization might be better served to focus on 3-5 ML3 process areas first - and should probably first focus on those where they perceive the greatest pain reduction or additional benefit.  I'll contend that the typical ML1 organization would be far better served to focus on 2-3 ML2 process areas rather than forming 7 groups, which just HAPPEN to have the same names as each o the ML2 process areas!

                                          I've never been a big believer in trying to achieve maturity levels as quickly as possible.  I would much prefer that the organization focus on gaining the benefits for the projects in a sustainable way than gaining another certificate for the trophy case.  After all, certifications don't produce software - people and projects do.

                                          Regards,

                                          Pat

                                           
                                           
                                           

                                          ----- Original Message -----
                                           
                                          From:  R C Goyal <mailto:rcgoyal@...>  
                                           
                                          To: cmmi_process_improvement@yahoogroups.com  
                                           
                                          Sent: Monday, April 30, 2007 10:54  PM
                                           
                                          Subject: Re: {Disarmed} Re: [CMMi Process  Improvement] Skipping ML2
                                           

                                           
                                           

                                          Dear Orhan

                                          Without entering into lengthy discussions may I  mention:
                                           
                                          • Measure about process being counterproductive is not easy. It may also  be one's own view point depending upon one's own expectation  
                                          • Again I would say going from ML1 to ML3  does not mean skipping  ML2, however it would also depend the meticulous gap-analysis and  improvements for correctly satisfying goals of ML2-PAs and to great extent  would depend on  quality of Class C (& B) appraisal format and  earnest execution.  
                                          • In my opinion you can not  satisfy ML3-PA's-goals (how well  institutionalized) without ML2 (what)
                                          Please comment on proper  alignment of my thought process?

                                          I always say that a model is means to  achieve improvement (even excellence) and the model is not the destination.  

                                          By  deciding  to go CMMI way , you got some directions for  your journey (e.g. L5 which is optimizing a indication of continual  process)

                                          Regards Goyal
                                           
                                           

                                           
                                           

                                           

                                           

                                          From: cmmi_process_improvement@yahoogroups.com  [ mailto:cmmi_ process_improvement@yahoogroups.com <mailto:process_improvement@yahoogroups.com> ]  On Behalf Of Orhan KALAYCI
                                          Sent: Monday, April 30, 2007  9:59 AM
                                          To: cmmi_process_improvement@yahoogroups.com
                                          Subject:  {Disarmed} Re: [CMMi Process Improvement] Skipping ML2

                                           
                                           
                                           

                                          Dear Goyal,

                                          This is a good question.  And In fact, this  question explains why we are (or I am alone) debating about it.

                                          I  am talking about this subject over and over because the importance of not  skipping levels is widely over looked.  
                                          Going from  ML1 to ML3 directly is counterproductive. I believe there are  a lot of companies at ML1 jump ML3.  Although these companies  certified as ML3, their productivity do not increase even decrease. Then it looks  like CMMI is counterproductive. What is worse, people do not  know that this counter productivity is a result of  skipping levels but they think CMMI makes these companies  less productive.  

                                          May the peace be upon you,
                                          Orhan
                                          www.xpi.ca <http://www.xpi.ca>
                                          www.nitelik.net <http://www.nitelik.net>


                                           
                                          2007/4/30, R C Goyal <  rcgoyal@... <mailto:rcgoyal@...> >:  

                                           
                                           
                                           
                                           

                                          Actually when an organization is being appraised for  higher level the appraisal is in any case being done for satisfying  goals of all PAs even for lower levels. Then why this subject of skipping is  being debated at all.



                                           
                                          On 4/29/07, Orhan  KALAYCI <  orhan.kalayci@xpi.ca <mailto:orhan.kalayci@...> > wrote:  

                                           
                                           
                                           
                                           

                                          Hello,

                                          I have difficulty in understanding why the most  experienced CMMI gurus do not support my proposal:  Let us not  encourage organizations to skip ML2!  

                                          This is counter  productive!  Do you really believe in it?  Do you support  not skipping MLs especially ML2?  (We see nobody targeting  directly ML5, it is ML2, the most skipped)

                                          My  questions:
                                          -  What value do you think ML2 provides  organizations with?
                                          -  Do you think skipping ML2 is really  counterproductive?  What kind of counter-productiveness does  it create?
                                          -  How can an organization understand it is ready  to start ML3 activities?  Not skipping ML2?

                                          May the peace  be upon you,
                                          Orhan
                                          www.xpi.ca <http://www.xpi.ca>
                                          www.nitelik.net <http://www.nitelik.net>

                                          You know my  analogy:  "Before federal and state laws (standard processes)  take over, we need to let sheriffs (PMs) to rule their counties  (Projects) for a while.  We know, this approach has been proved  to be successful in USA.  When ML2 tried to be skipped in Soviet  Unions, where they rushed to standardize everything to catch the  economics in scale.  We now know this approach (skipping ML2) is  not successful, on the opposite it is really counterproductive.   At first sight, it looks like standards should provide productivity,  because everybody will work in the same way, synchronization,   harmony, etc will be established.  However, if you rush to  standardize too quick.  It results in just the opposite.  As  I keep saying, life is full of contradictions."




                                        • Patrick OToole
                                          Orhan, On THIS post we are in TOTAL agreement! Regards, Pat ... From: Orhan KALAYCI To: cmmi_process_improvement@yahoogroups.com Sent: Thursday, May 03, 2007
                                          Message 20 of 30 , May 3, 2007
                                            
                                             
                                            Orhan,
                                             
                                            On THIS post we are in TOTAL agreement!
                                             
                                            Regards,
                                             
                                            Pat
                                             
                                            ----- Original Message -----
                                            Sent: Thursday, May 03, 2007 9:49 AM
                                            Subject: Re: {Disarmed} Re: [CMMi Process Improvement] Skipping ML2

                                            Hi group, how are you?

                                            Thanks to these posts. I believe it is now clear that there are two dimensions of skipping levels. 
                                            From ML1 -> ML3 or from CL1 -> CL3.

                                            When we say ML1->ML3 in fact this is a short way of saying a set of PAs going from CL1->CL3. 
                                            If fact, when you zoom into this situation real problem is going from CL1->CL3.  So if an PA is already at CL2, of course there is no problem for this specific PA to continue towards CL3.  However, be aware of the problem of synchronization among other PAs that may at CL2. 

                                            Let us not forget the spirit:
                                            Let us not encourage people writing processes in fact when they do not know these processes. In other words, when they are not ready to write the standard processes for others to follow.  First, they should discover the processes by not reading the books about them but living in these processes for a while (ML2).  At ML2, we just name the processes and ask people to develop their own understanding by try it live it and learn it.  

                                            May the peace be upon you,
                                            Orhan
                                            www.xpi.ca
                                            www.nitelik. net

                                            2007/5/3, Patrick OToole < PACT.otoole@ att.net>:

                                             
                                            Everyone,
                                             
                                            Here's another "problem" that I'm having with the "don't do things at ML3 until you've got ML2 nailed" position.  Let's say an organization has staffed its process improvement project such that it can focus on improving 7 things concurrently.  Let's say they initially form one team to work on REQM, another to work on PP, another to work on PMC, etc.
                                             
                                            But, in this organization, project management is already in pretty good shape, while MA and PPQA are pretty much starting from scratch.  In such a case, the PP and PMC teams may wrap up pretty quickly - within a month or two they may fill the couple of gaps, train the other PMs to adopt the new practices in a value-added way and they're done.
                                             
                                            In the meantime, the MA and PPQA teams are trying to understand what the associated practices are all about, they are off getting trained on their respective process areas, and have put together a one-year plan to become "model compliant."
                                             
                                            Since the organization is already staffed to run 7 concurrent improvement teams, they COULD re-allocate all the PP and PMC folks to assist with MA and PPQA, and they'd wind up with twice as many people in a group grope.  The PP and PMC teams would now need to be trained and would be playing 2 months of catch up.  This MAY reduce the 12 month schedule by a month or two - or it may not.
                                             
                                            Alternatively, the PP and PMC teams could be re-directed to work on RSKM and IPM, maybe targeting CL1 or CL2 in these two process areas.  After all, this is simply an extension of the project management work they have already performed AND there may be a pretty solid foundation in some of these practices as well.
                                             
                                            Would the organization be WRONG if they decided to go with the second option?
                                             
                                            And, if they get IPM and RSKM to CL2 within 6 months, would they be wrong to then start working on PP and PMC GP3.1 and GP3.2?  Is there anything wrong with trying to standardize the practices and processes that were pretty much deeply institutionalized anyway?
                                             
                                            As much as I agree with the basic principle that we "shouldn't skip maturity levels," I'm merely suggesting that it is hardly ever that simple, and that we really need to look at each unique situation to determine what's best for THAT organization.  Just as there is no one way to implement any given CMMI practice, and we encourage organizations to implement the practices in a way that best fits their needs, shouldn't we be somewhat flexible in our statements about the ORDER in which they implement the practices?
                                             
                                            And on it goes...
                                             
                                            Regards,
                                             
                                            Pat
                                             
                                             
                                             
                                             
                                            ----- Original Message -----
                                            Sent: Wednesday, May 02, 2007 1:29 PM
                                            Subject: Re: {Disarmed} Re: [CMMi Process Improvement] Skipping ML2

                                            Hello group, how are you?

                                            To be able discuss or debate on ML2, ML3 or CL2, CL3.  Let me define what they mean:
                                            First of all, an ML is not equal to a set of PAs i.e. ML <> {PAs}

                                            CMMI has a spirit, so do each ML and CL. 

                                            ML2 means "Project management" is in place including processes (tasks) for process areas of ML3 in the project plans. However, they all are at the level of CL2.  We can never think about a project plan without validation or verification tasks in it.  So at ML2, it is for sure there are processes representing VER and VAL PAs (at CL2).  However they are not tailored from set of organizational standard processes (CL3).  Let me continue, can you think of a real project plan without processes representing TS, PI, RD?  No, it is nonsense to have a project plan, without requirements analysis or developing a solution (design, coding, etc).  Is there a chance that we can deliver our product to customer without integrating product components?  At ML2, they (processes (tasks) for PAs of ML2 + ML3) all should be there but they are all at CL2.  Spirit of ML3 is not a set of PAs.  What makes an organization at ML3 (spirit) or a process at CL3 (spirit)  is GG3, i.e., standard processes and tailoring guideline.  Management of standard processes is OPD. 

                                            OPF should belong ML2.  I will write a CR for that. 
                                            In fact, I need to write another CR about making it "explicit" that ML2 = PAs of (ML2+ML3) should be at CL2.  This is now implicit in PP in WBS, Work packages, task descriptions, etc. 

                                            ML2 and CL2 is about "living" the processes. 
                                            ML3 and CL3 is about "writing" the processes.

                                            So, do not start "writing" the processes (tasks) before "living" them.  If you do so, it is highly probable that you will write processes (tasks) unlivable.  And (kudos to Jeff) these processes unfortunately will not set you free, just the opposite.  They will not bring you joy and freedom but misery and slavery.  

                                            May the peace and the freedom be upon you,
                                            Orhan
                                            www.xpi.ca
                                            www.nitelik. net


                                            2007/5/1, Patrick OToole < PACT.otoole@ att.net>:

                                             
                                            Ah, I think I'm finally getting a handle on this discussion.  It seems to me there are two concurrent thought streams running through this dialogue (or at least through my head) - and it all comes down to the phrase "skipping maturity levels."
                                             
                                            One interpretation is "let's work on all the ML2 and ML3 process areas at the same time - doing ML2 first, and then working on ML3 afterwards is wimpy - real men do it all!"
                                             
                                            The other interpretation is "let's work on SOME of the ML3 things before we work on SOME of the ML2 things.  Afterall, we have project pain that would gain some relief from enhancing our Requirements Elicitation and/or Peer Review processes, and I know that these are ML3 things and that we're an ML1 organization, but still..."
                                             
                                            I would STRONGLY agree that typical ML1 organizations are NOT well-served to follow the first interpretation - let's do everything all at once.  Stabilize the project domain before you attack the organizational issues.  Get some credibility and momentum established before you start trying to conquer the world.  (I'll still contend that SOME ML1 organizations, who already have MANY of the ML2 process areas well established are probably in a position where they can start working on some of the ML3 process areas while they are finishing up the ML2 stuff, but that's not what I mean by a "typical ML1 organization.")
                                             
                                            Most of my earlier posts were focused on the second interpretation of "skipping maturity levels."  I am certainly NOT a purist when it comes to establishing infrastructure and fixing pain.  If an SEPG-like thing is going to help us get focus and momentum for process improvement, then let's dip into OPF and figure out what an SEPG is and what kinds of things they do.  If OPD can give us a bit of guidance on establishing life cycles, how to write process documentation, and establishing a measurement repository, then let's take a look at that as well.
                                             
                                            For pain relief - if improving the peer review process is going to eliminate some pain that we feel in unit, integration, and system testing, then I'm all for it.  If introducing JAD sessions is going to get us better requirements, then I'm all for that too. 
                                             
                                            I do not adhere to the philosophy that "you must ONLY do ML2 stuff before you work on any of the ML3 stuff" - and if THAT's what someone means by "skipping maturity levels" then they'd better be prepared for a vigorous debate. 
                                             
                                            We must also keep in mind, as I've mentioned before, that organizations may already have a bunch of good stuff in place when they first pick up the model.  They may already have a Project Management Office that has some good risk management practices and may be running IPM-like integrated programs.  They may have an independent test group that is doing their portion of VER extremely well.  They may have a separate training organization that is already outstanding at many of the OT practices.  I would NOT tell these organizations to "quit wasting their time improving these ML3 like practices because there are some gaps at ML2."
                                             
                                            One final thought - with 11 process areas, I would contend that ML3 is way too much for a typical ML2 organization to bite off all at once.  The typical ML2 organization might be better served to focus on 3-5 ML3 process areas first - and should probably first focus on those where they perceive the greatest pain reduction or additional benefit.  I'll contend that the typical ML1 organization would be far better served to focus on 2-3 ML2 process areas rather than forming 7 groups, which just HAPPEN to have the same names as each o the ML2 process areas!
                                             
                                            I've never been a big believer in trying to achieve maturity levels as quickly as possible.  I would much prefer that the organization focus on gaining the benefits for the projects in a sustainable way than gaining another certificate for the trophy case.  After all, certifications don't produce software - people and projects do.
                                             
                                            Regards,
                                             
                                            Pat
                                             
                                             
                                             
                                             
                                            ----- Original Message -----
                                            From: R C Goyal
                                            Sent: Monday, April 30, 2007 10:54 PM
                                            Subject: Re: {Disarmed} Re: [CMMi Process Improvement] Skipping ML2

                                            Dear Orhan

                                            Without entering into lengthy discussions may I mention:

                                            • Measure about process being counterproductive is not easy. It may also be one's own view point depending upon one's own expectation
                                            • Again I would say going from ML1 to ML3  does not mean skipping ML2, however it would also depend the meticulous gap-analysis and improvements for correctly satisfying goals of ML2-PAs and to great extent would depend on  quality of Class C (& B) appraisal format and earnest execution.
                                            • In my opinion you can not  satisfy ML3-PA's-goals (how well institutionalized) without ML2 (what)
                                            Please comment on proper alignment of my thought process?

                                            I always say that a model is means to achieve improvement (even excellence) and the model is not the destination.

                                            By  deciding  to go CMMI way , you got some directions for your journey (e.g. L5 which is optimizing a indication of continual process)

                                            Regards Goyal


                                            From: cmmi_process_ improvement@ yahoogroups. com [mailto:cmmi_ process_improvement @yahoogroups. com] On Behalf Of Orhan KALAYCI
                                            Sent: Monday, April 30, 2007 9:59 AM
                                            To: cmmi_process_ improvement@ yahoogroups. com
                                            Subject: {Disarmed} Re: [CMMi Process Improvement] Skipping ML2

                                            Dear Goyal,

                                            This is a good question.  And In fact, this question explains why we are (or I am alone) debating about it.

                                            I am talking about this subject over and over because the importance of not skipping levels is widely over looked.  Going from ML1 to ML3 directly is counterproductive.  I believe there are a lot of companies at ML1 jump ML3.  Although these companies certified as ML3, their productivity do not increase even decrease. Then it looks like CMMI is counterproductive.  What is worse, people do not know that this counter productivity is a result of skipping levels but they think CMMI makes these companies less productive. 

                                            May the peace be upon you,
                                            Orhan
                                            www.xpi.ca
                                            www.nitelik. net


                                            2007/4/30, R C Goyal < rcgoyal@gmail. com>:

                                            Actually when an organization is being appraised for higher level the appraisal is in any case being done for satisfying goals of all PAs even for lower levels. Then why this subject of skipping is being debated at all.



                                            On 4/29/07, Orhan KALAYCI < orhan.kalayci@ xpi.ca> wrote:

                                            Hello,

                                            I have difficulty in understanding why the most experienced CMMI gurus do not support my proposal:  Let us not encourage organizations to skip ML2! 

                                            This is counter productive!  Do you really believe in it?  Do you support not skipping MLs especially ML2?  (We see nobody targeting directly ML5, it is ML2, the most skipped)

                                            My questions:
                                            -  What value do you think ML2 provides organizations with?
                                            -  Do you think skipping ML2 is really counterproductive?  What kind of counter-productiven ess does it create?
                                            -  How can an organization understand it is ready to start ML3 activities?  Not skipping ML2?

                                            May the peace be upon you,
                                            Orhan
                                            www.xpi.ca
                                            www.nitelik. net

                                            You know my analogy:  "Before federal and state laws (standard processes) take over, we need to let sheriffs (PMs) to rule their counties (Projects) for a while.  We know, this approach has been proved to be successful in USA.  When ML2 tried to be skipped in Soviet Unions, where they rushed to standardize everything to catch the economics in scale.  We now know this approach (skipping ML2) is not successful, on the opposite it is really counterproductive.  At first sight, it looks like standards should provide productivity, because everybody will work in the same way, synchronization,  harmony, etc will be established.  However, if you rush to standardize too quick.  It results in just the opposite.  As I keep saying, life is full of contradictions."




                                            --
                                            Thanks & Regards R C Goyal
                                            Management Committee  Member,  PMI & CSI Mumbai Chapter
                                            BE AN AMBASSADOR OF QUALITY
                                            programs@pmimumbaic hapter.org , rcgoyal@gmail. com; rcgoyal@avrms. com;
                                            Mob: 00919869463964; 00912267108110





                                            --
                                            Thanks & Regards R C Goyal
                                            Management Committee  Member,  PMI & CSI Mumbai Chapter
                                            BE AN AMBASSADOR OF QUALITY
                                            programs@pmimumbaic hapter.org, rcgoyal@gmail. com; rcgoyal@avrms. com;
                                            Mob: 00919869463964; 00912267108110



                                          • Patrick OToole
                                            Orhan, And here, we are NOT in total agreement. Just because the CMMI ALLOWS ML2 organizations to have different project-level processes, doesn t mean that you
                                            Message 21 of 30 , May 3, 2007
                                              
                                               
                                              Orhan,
                                               
                                              And here, we are NOT in total agreement. 
                                               
                                              Just because the CMMI ALLOWS ML2 organizations to have different project-level processes, doesn't mean that you HAVE to do it this way.  If an ML1 organization already has standard processes that are working well for them, and management wants to insist that these processes continue to be used and tailored appropriately, then I see nothing wrong with that - nor, apparently, does the continuous representation.
                                               
                                              An organization could certainly have a capability profile like:
                                               
                                              REQM: CL2
                                              PP: CL3
                                              PMC: CL3
                                              SAM: CL1
                                              MA: CL2
                                              PPQA: CL1
                                              CM: CL3
                                              IPM: CL2
                                              RSKM: CL3
                                               
                                              They are CLEARLY ML1 equivalent, but if they're using the continuous representation then they may not give a hoot about their maturity level, or maturity level equivalence.  They are simply focusing on establishing standard processes where they perceive their projects will achieve the most benefit.  They may NEVER decide to improve PPQA or SAM - but they may decide that CM and PP warrant quantitative management and take them to CL4.
                                               
                                              I would never tell this organization that each project should feel free to forego use of their standard PP, PMC, CM, and RSKM practices simply because they are ML1 equivalent!
                                               
                                              If we were going to do that, then why have the continuous representation at all?
                                               
                                              Regards,
                                               
                                              Pat
                                               
                                               
                                              ----- Original Message -----
                                              Sent: Thursday, May 03, 2007 9:36 AM
                                              Subject: Re: {Disarmed} Re: [CMMi Process Improvement] Skipping ML2

                                              Pat,

                                              No, I don't think they need to abandon anything already done.  But I believe it is better to let people at a ML1 org. know they have right to re-write these documents according to needs of their projects.  In other words, tailoring in this organization (ML1) should provide projects with really huge space to adapt these docs to their specific needs.

                                              May the peace be upon you,
                                              Orhan
                                              www.xpi.ca
                                              www.nitelik. net


                                              2007/5/3, Patrick OToole < PACT.otoole@ att.net>:

                                               
                                              Orhan,
                                               
                                              If an ML1 organization already has documented processes for its software and test engineering activities, and these documented processes include ways to tailor them to best meet their needs, should they abandon them until they completely implement all practices associated with ML2?
                                               
                                              Regards,
                                               
                                              Pat
                                               
                                               
                                              ----- Original Message -----
                                              Sent: Wednesday, May 02, 2007 1:29 PM
                                              Subject: Re: {Disarmed} Re: [CMMi Process Improvement] Skipping ML2

                                              Hello group, how are you?

                                              To be able discuss or debate on ML2, ML3 or CL2, CL3.  Let me define what they mean:
                                              First of all, an ML is not equal to a set of PAs i.e. ML <> {PAs}

                                              CMMI has a spirit, so do each ML and CL. 

                                              ML2 means "Project management" is in place including processes (tasks) for process areas of ML3 in the project plans. However, they all are at the level of CL2.  We can never think about a project plan without validation or verification tasks in it.  So at ML2, it is for sure there are processes representing VER and VAL PAs (at CL2).  However they are not tailored from set of organizational standard processes (CL3).  Let me continue, can you think of a real project plan without processes representing TS, PI, RD?  No, it is nonsense to have a project plan, without requirements analysis or developing a solution (design, coding, etc).  Is there a chance that we can deliver our product to customer without integrating product components?  At ML2, they (processes (tasks) for PAs of ML2 + ML3) all should be there but they are all at CL2.  Spirit of ML3 is not a set of PAs.  What makes an organization at ML3 (spirit) or a process at CL3 (spirit)  is GG3, i.e., standard processes and tailoring guideline.  Management of standard processes is OPD. 

                                              OPF should belong ML2.  I will write a CR for that. 
                                              In fact, I need to write another CR about making it "explicit" that ML2 = PAs of (ML2+ML3) should be at CL2.  This is now implicit in PP in WBS, Work packages, task descriptions, etc. 

                                              ML2 and CL2 is about "living" the processes. 
                                              ML3 and CL3 is about "writing" the processes.

                                              So, do not start "writing" the processes (tasks) before "living" them.  If you do so, it is highly probable that you will write processes (tasks) unlivable.  And (kudos to Jeff) these processes unfortunately will not set you free, just the opposite.  They will not bring you joy and freedom but misery and slavery.  

                                              May the peace and the freedom be upon you,
                                              Orhan
                                              www.xpi.ca
                                              www.nitelik. net


                                              2007/5/1, Patrick OToole < PACT.otoole@ att.net>:

                                               
                                              Ah, I think I'm finally getting a handle on this discussion.  It seems to me there are two concurrent thought streams running through this dialogue (or at least through my head) - and it all comes down to the phrase "skipping maturity levels."
                                               
                                              One interpretation is "let's work on all the ML2 and ML3 process areas at the same time - doing ML2 first, and then working on ML3 afterwards is wimpy - real men do it all!"
                                               
                                              The other interpretation is "let's work on SOME of the ML3 things before we work on SOME of the ML2 things.  Afterall, we have project pain that would gain some relief from enhancing our Requirements Elicitation and/or Peer Review processes, and I know that these are ML3 things and that we're an ML1 organization, but still..."
                                               
                                              I would STRONGLY agree that typical ML1 organizations are NOT well-served to follow the first interpretation - let's do everything all at once.  Stabilize the project domain before you attack the organizational issues.  Get some credibility and momentum established before you start trying to conquer the world.  (I'll still contend that SOME ML1 organizations, who already have MANY of the ML2 process areas well established are probably in a position where they can start working on some of the ML3 process areas while they are finishing up the ML2 stuff, but that's not what I mean by a "typical ML1 organization.")
                                               
                                              Most of my earlier posts were focused on the second interpretation of "skipping maturity levels."  I am certainly NOT a purist when it comes to establishing infrastructure and fixing pain.  If an SEPG-like thing is going to help us get focus and momentum for process improvement, then let's dip into OPF and figure out what an SEPG is and what kinds of things they do.  If OPD can give us a bit of guidance on establishing life cycles, how to write process documentation, and establishing a measurement repository, then let's take a look at that as well.
                                               
                                              For pain relief - if improving the peer review process is going to eliminate some pain that we feel in unit, integration, and system testing, then I'm all for it.  If introducing JAD sessions is going to get us better requirements, then I'm all for that too. 
                                               
                                              I do not adhere to the philosophy that "you must ONLY do ML2 stuff before you work on any of the ML3 stuff" - and if THAT's what someone means by "skipping maturity levels" then they'd better be prepared for a vigorous debate. 
                                               
                                              We must also keep in mind, as I've mentioned before, that organizations may already have a bunch of good stuff in place when they first pick up the model.  They may already have a Project Management Office that has some good risk management practices and may be running IPM-like integrated programs.  They may have an independent test group that is doing their portion of VER extremely well.  They may have a separate training organization that is already outstanding at many of the OT practices.  I would NOT tell these organizations to "quit wasting their time improving these ML3 like practices because there are some gaps at ML2."
                                               
                                              One final thought - with 11 process areas, I would contend that ML3 is way too much for a typical ML2 organization to bite off all at once.  The typical ML2 organization might be better served to focus on 3-5 ML3 process areas first - and should probably first focus on those where they perceive the greatest pain reduction or additional benefit.  I'll contend that the typical ML1 organization would be far better served to focus on 2-3 ML2 process areas rather than forming 7 groups, which just HAPPEN to have the same names as each o the ML2 process areas!
                                               
                                              I've never been a big believer in trying to achieve maturity levels as quickly as possible.  I would much prefer that the organization focus on gaining the benefits for the projects in a sustainable way than gaining another certificate for the trophy case.  After all, certifications don't produce software - people and projects do.
                                               
                                              Regards,
                                               
                                              Pat
                                               
                                               
                                               
                                               
                                              ----- Original Message -----
                                              From: R C Goyal
                                              Sent: Monday, April 30, 2007 10:54 PM
                                              Subject: Re: {Disarmed} Re: [CMMi Process Improvement] Skipping ML2

                                              Dear Orhan

                                              Without entering into lengthy discussions may I mention:

                                              • Measure about process being counterproductive is not easy. It may also be one's own view point depending upon one's own expectation
                                              • Again I would say going from ML1 to ML3  does not mean skipping ML2, however it would also depend the meticulous gap-analysis and improvements for correctly satisfying goals of ML2-PAs and to great extent would depend on  quality of Class C (& B) appraisal format and earnest execution.
                                              • In my opinion you can not  satisfy ML3-PA's-goals (how well institutionalized) without ML2 (what)
                                              Please comment on proper alignment of my thought process?

                                              I always say that a model is means to achieve improvement (even excellence) and the model is not the destination.

                                              By  deciding  to go CMMI way , you got some directions for your journey (e.g. L5 which is optimizing a indication of continual process)

                                              Regards Goyal


                                              From: cmmi_process_ improvement@ yahoogroups. com [mailto:cmmi_ process_improvement @yahoogroups. com] On Behalf Of Orhan KALAYCI
                                              Sent: Monday, April 30, 2007 9:59 AM
                                              To: cmmi_process_ improvement@ yahoogroups. com
                                              Subject: {Disarmed} Re: [CMMi Process Improvement] Skipping ML2

                                              Dear Goyal,

                                              This is a good question.  And In fact, this question explains why we are (or I am alone) debating about it.

                                              I am talking about this subject over and over because the importance of not skipping levels is widely over looked.  Going from ML1 to ML3 directly is counterproductive.  I believe there are a lot of companies at ML1 jump ML3.  Although these companies certified as ML3, their productivity do not increase even decrease. Then it looks like CMMI is counterproductive.  What is worse, people do not know that this counter productivity is a result of skipping levels but they think CMMI makes these companies less productive. 

                                              May the peace be upon you,
                                              Orhan
                                              www.xpi.ca
                                              www.nitelik. net


                                              2007/4/30, R C Goyal < rcgoyal@gmail. com>:

                                              Actually when an organization is being appraised for higher level the appraisal is in any case being done for satisfying goals of all PAs even for lower levels. Then why this subject of skipping is being debated at all.



                                              On 4/29/07, Orhan KALAYCI < orhan.kalayci@ xpi.ca> wrote:

                                              Hello,

                                              I have difficulty in understanding why the most experienced CMMI gurus do not support my proposal:  Let us not encourage organizations to skip ML2! 

                                              This is counter productive!  Do you really believe in it?  Do you support not skipping MLs especially ML2?  (We see nobody targeting directly ML5, it is ML2, the most skipped)

                                              My questions:
                                              -  What value do you think ML2 provides organizations with?
                                              -  Do you think skipping ML2 is really counterproductive?  What kind of counter-productiven ess does it create?
                                              -  How can an organization understand it is ready to start ML3 activities?  Not skipping ML2?

                                              May the peace be upon you,
                                              Orhan
                                              www.xpi.ca
                                              www.nitelik. net

                                              You know my analogy:  "Before federal and state laws (standard processes) take over, we need to let sheriffs (PMs) to rule their counties (Projects) for a while.  We know, this approach has been proved to be successful in USA.  When ML2 tried to be skipped in Soviet Unions, where they rushed to standardize everything to catch the economics in scale.  We now know this approach (skipping ML2) is not successful, on the opposite it is really counterproductive.  At first sight, it looks like standards should provide productivity, because everybody will work in the same way, synchronization,  harmony, etc will be established.  However, if you rush to standardize too quick.  It results in just the opposite.  As I keep saying, life is full of contradictions."




                                              --
                                              Thanks & Regards R C Goyal
                                              Management Committee  Member,  PMI & CSI Mumbai Chapter
                                              BE AN AMBASSADOR OF QUALITY
                                              programs@pmimumbaic hapter.org , rcgoyal@gmail. com; rcgoyal@avrms. com;
                                              Mob: 00919869463964; 00912267108110





                                              --
                                              Thanks & Regards R C Goyal
                                              Management Committee  Member,  PMI & CSI Mumbai Chapter
                                              BE AN AMBASSADOR OF QUALITY
                                              programs@pmimumbaic hapter.org, rcgoyal@gmail. com; rcgoyal@avrms. com;
                                              Mob: 00919869463964; 00912267108110



                                            • Andre Heijstek
                                              Hi Orhan, I m afraid I still have to disagree. ... The spirit of TS is that systems are being developed. Systems are already developed in ML1 and even in CL0
                                              Message 22 of 30 , May 3, 2007
                                                Hi Orhan,

                                                I'm afraid I still have to disagree.

                                                On 3-mei-2007, at 21:20, Orhan KALAYCI wrote:

                                                OPF:  The spirit of OPF is setting targets for process improvement and managing these targets.  Setting process improvement targets begins with ML2.  So it should be placed in ML2. 

                                                The spirit of TS is that systems are being developed. Systems are already developed in ML1 and even in CL0 processes. So in your line of reasoning TS (and RS, and PI, and VER/VAL) should be moved to ML1!
                                                If you move everything to the lowest level - using the argument that the spirit of this PA is being performed always - there will be no levels anymore, and thus no guidance.

                                                Regards, Andre.
                                              • Patrick OToole
                                                Andre, One of my absolutely favorite quotes from one of your posts was when you said, The maturity levels weren t determined by what organizations should DO
                                                Message 23 of 30 , May 3, 2007
                                                   
                                                  Andre,
                                                   
                                                  One of my absolutely favorite quotes from one of your posts was when you said, "The maturity levels weren't determined by what organizations should DO first - rather, the maturity levels were determined by what organizations should IMPROVE first."
                                                   
                                                  After all, organizations gather requirements, produce designs and code, and typically test their products at ML1.  However, the model would suggest that it is more beneficial to get a fundamental handle on requirements change over the life of the project, get fundamental control of the configuration items, establish basic project management practices, etc.
                                                   
                                                  It's hard to get momentum on software or process engineering in the midst of chaos.  ML2 is all about project triage - stabilize the project and THEN the organization will have the proper environment to effect real and lasting change.
                                                   
                                                  Regards,
                                                   
                                                  Pat
                                                   
                                                   
                                                   
                                                  ----- Original Message -----
                                                  Sent: Friday, May 04, 2007 12:59 AM
                                                  Subject: Re: {Disarmed} Re: [CMMi Process Improvement] Skipping ML2

                                                  Hi Orhan,


                                                  I'm afraid I still have to disagree.

                                                  On 3-mei-2007, at 21:20, Orhan KALAYCI wrote:

                                                  OPF:  The spirit of OPF is setting targets for process improvement and managing these targets.  Setting process improvement targets begins with ML2.  So it should be placed in ML2. 

                                                  The spirit of TS is that systems are being developed. Systems are already developed in ML1 and even in CL0 processes. So in your line of reasoning TS (and RS, and PI, and VER/VAL) should be moved to ML1!
                                                  If you move everything to the lowest level - using the argument that the spirit of this PA is being performed always - there will be no levels anymore, and thus no guidance.

                                                  Regards, Andre.

                                                • Bruce Epstein
                                                  Pat, Actually, I believe there s a THIRD thought stream running through this discussion (as well as a variant on your second thought stream): Third
                                                  Message 24 of 30 , May 4, 2007
                                                    Pat,

                                                    Actually, I believe there's a THIRD thought stream running through
                                                    this discussion (as well as a variant on your second thought stream):

                                                    Third interpretation: An ML1 organisation sets out to work on some
                                                    set of process areas (not necessarily congruent to any particular
                                                    ML) by defining a set of "organisational standard processes". As
                                                    has already been pointed out, this approach CAN work if there are
                                                    pre-existing institutionalised processes somewhere that are being
                                                    generalised to the entire organisation. This approach is doomed to
                                                    failure, however, if the "new" processes are defined disconnected
                                                    from the needs of the real projects!

                                                    True story: one IT director recently decided that to avoid the type
                                                    of chaos that Andre described in a recent post, ALL projects must
                                                    follow an identical process. A noble goal, but unfortunately, the
                                                    process was defined so rigidly that NO projects were able to
                                                    actually apply it!

                                                    Variant on your second theme: sometimes (dare I say often?), an ML1
                                                    organisation discovers that it MUST work on some ML3 concepts in
                                                    order to satisfy ML2. For example, PP SP1.3 requires projects to
                                                    define their life cycle. Well, if the existing lifecycle is no
                                                    better defined than
                                                    1. receive a request
                                                    2. do some technical stuff
                                                    3. deliver the product
                                                    then it is absolutely necessary to work on some of the concepts of
                                                    RD, TS, PI, VER and VAL in parallel to, or even before, work on PP.
                                                    And since we probably don't want every project inventing its own
                                                    lifecycle, we'll try to define a limited number of acceptable
                                                    lifecycles at the organisational level (hmm, this starts to sound
                                                    like OPD, but implemented in a way that reflects the lower maturity
                                                    level of the organisation).

                                                    Final thought: Perhaps it's just my bad karma, but most of my
                                                    clients over the past few years have started from a point quite far
                                                    away from ML2, where none of the 7 PAs of ML2 would have satisfied
                                                    even CL1 had they been appraised. I suspect that Orhan has faced
                                                    similar situations. This is very different from the "almost ML2"
                                                    starting points that you describe below.

                                                    Regards,
                                                    Bruce

                                                    --- In cmmi_process_improvement@yahoogroups.com, "Patrick OToole"
                                                    <PACT.otoole@...> wrote:
                                                    >
                                                    >
                                                    > Ah, I think I'm finally getting a handle on this discussion. It
                                                    seems to me there are two concurrent thought streams running through
                                                    this dialogue (or at least through my head) - and it all comes down
                                                    to the phrase "skipping maturity levels."
                                                    >
                                                    > One interpretation is "let's work on all the ML2 and ML3 process
                                                    areas at the same time - doing ML2 first, and then working on ML3
                                                    afterwards is wimpy - real men do it all!"
                                                    >
                                                    > The other interpretation is "let's work on SOME of the ML3 things
                                                    before we work on SOME of the ML2 things. Afterall, we have project
                                                    pain that would gain some relief from enhancing our Requirements
                                                    Elicitation and/or Peer Review processes, and I know that these are
                                                    ML3 things and that we're an ML1 organization, but still..."
                                                    >
                                                    > I would STRONGLY agree that typical ML1 organizations are NOT well-
                                                    served to follow the first interpretation - let's do everything all
                                                    at once. Stabilize the project domain before you attack the
                                                    organizational issues. Get some credibility and momentum
                                                    established before you start trying to conquer the world. (I'll
                                                    still contend that SOME ML1 organizations, who already have MANY of
                                                    the ML2 process areas well established are probably in a position
                                                    where they can start working on some of the ML3 process areas while
                                                    they are finishing up the ML2 stuff, but that's not what I mean by
                                                    a "typical ML1 organization.")
                                                    >
                                                    > Most of my earlier posts were focused on the second interpretation
                                                    of "skipping maturity levels." I am certainly NOT a purist when it
                                                    comes to establishing infrastructure and fixing pain. If an SEPG-
                                                    like thing is going to help us get focus and momentum for process
                                                    improvement, then let's dip into OPF and figure out what an SEPG is
                                                    and what kinds of things they do. If OPD can give us a bit of
                                                    guidance on establishing life cycles, how to write process
                                                    documentation, and establishing a measurement repository, then let's
                                                    take a look at that as well.
                                                    >
                                                    > For pain relief - if improving the peer review process is going to
                                                    eliminate some pain that we feel in unit, integration, and system
                                                    testing, then I'm all for it. If introducing JAD sessions is going
                                                    to get us better requirements, then I'm all for that too.
                                                    >
                                                    > I do not adhere to the philosophy that "you must ONLY do ML2 stuff
                                                    before you work on any of the ML3 stuff" - and if THAT's what
                                                    someone means by "skipping maturity levels" then they'd better be
                                                    prepared for a vigorous debate.
                                                    >
                                                    > We must also keep in mind, as I've mentioned before, that
                                                    organizations may already have a bunch of good stuff in place when
                                                    they first pick up the model. They may already have a Project
                                                    Management Office that has some good risk management practices and
                                                    may be running IPM-like integrated programs. They may have an
                                                    independent test group that is doing their portion of VER extremely
                                                    well. They may have a separate training organization that is
                                                    already outstanding at many of the OT practices. I would NOT tell
                                                    these organizations to "quit wasting their time improving these ML3
                                                    like practices because there are some gaps at ML2."
                                                    >
                                                    > One final thought - with 11 process areas, I would contend that
                                                    ML3 is way too much for a typical ML2 organization to bite off all
                                                    at once. The typical ML2 organization might be better served to
                                                    focus on 3-5 ML3 process areas first - and should probably first
                                                    focus on those where they perceive the greatest pain reduction or
                                                    additional benefit. I'll contend that the typical ML1 organization
                                                    would be far better served to focus on 2-3 ML2 process areas rather
                                                    than forming 7 groups, which just HAPPEN to have the same names as
                                                    each o the ML2 process areas!
                                                    >
                                                    > I've never been a big believer in trying to achieve maturity
                                                    levels as quickly as possible. I would much prefer that the
                                                    organization focus on gaining the benefits for the projects in a
                                                    sustainable way than gaining another certificate for the trophy
                                                    case. After all, certifications don't produce software - people and
                                                    projects do.
                                                    >
                                                    > Regards,
                                                    >
                                                    > Pat
                                                    >
                                                  Your message has been successfully submitted and would be delivered to recipients shortly.