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

Re: [CMMi Process Improvement] PI practices enlightenment

Expand Messages
  • cloony.william
    Hi Pat, I Agree with your comments... One problem i see in this approach is, not only the appraisal go as Managed Discovery but also the maturity level is
    Message 1 of 42 , May 31, 2011
    • 0 Attachment
      Hi Pat,

      I Agree with your comments...

      One problem i see in this approach is, not only the appraisal go as "Managed Discovery" but also the maturity level is achieved by the organization as "Managed Maturity with Selected Projects"... This is because each time incrementally the organization fills the gap in the practices and at the end they achieve 100% Compliance..

      However if the projects taken in SCAMPI C/B or Readiness Review and in SCAMPI A are different then we can say the Organization has achieved the maturity.. But most of the time the projects are retained and taken to SCAMPI A...

      Eventhough this is a total deviation from the original post , I believe this is a valid concern ....

      Cloony

      --- In cmmi_process_improvement@yahoogroups.com, "Pat OToole" <PACT.otoole@...> wrote:
      >
      >
      > The trick with the SCAMPI A appraisal method (or any appraisal method for that matter), is to “clear the decks of the easy stuff.” That is, determine which practices are implemented in a matter that is CLEARLY aligned with the model. Get that stuff (hopefully in the 75% â€" 90% range) out of the way as quickly as possible so that the team can get focused on: (1) the stuff where the alignment is not as apparent; and/or (2) the stuff where the practice implementation is weaker.
      >
      > The lead appraiser should be encouraging the team that, once you’ve concluded a practice has strong evidence that indicates it’s implemented on the project, move on as quickly as possible to the next practice â€" and that’s the case whether you’re doing initial documentation review or interviews â€" use your time effectively.
      >
      > The same holds true for establishing PIIDs. If you have done a Class C or other appraisal in advance of the SCAMPI A, look at the evidence that led THAT appraisal team to conclude that the practice was in fine shape (or not). If it was deemed to be OK, then you’ll want to show similar evidence in the SCAMPI A (if it ain’t broke, don’t fix it!) If it was deemed weak in the initial appraisal event, then you’ll want to show the new and improved evidence that demonstrates the fact that you’ve taken remedial action and strengthened your implementation. So to reduce your PIIDing time â€" reuse as much as you can and simply supplement or replace when you can’t.
      >
      > One reason PIIDs take so blasted long to populate PIIDs is that organizations tend to OVER-populate them. Because they want to show that they are rock solid, they try to demonstrate 100%+ compliance â€" and if that takes 8-10 artifacts per practice per project, then so be it!
      >
      > To me, excessive PIIDing is not only a waste of YOUR time, it’s also a waste of the appraisal team’s time. I believe that 50-75% of practice implementation can be “proven” with a singular artifact. That is, PP SP1.1 on Project 1 should most likely take only 1 direct artifact to convince the appraisal team member of that practice’s implementation on that project. After all, if it convinced YOU that you’re doing it, and the appraisal team is simply validating your own conclusions, then 1 should be enough.
      >
      > More than half of the remaining practices should only need 2 artifacts, and the final practices should need, at most, 3. If it takes more than 3 artifacts to demonstrate that a practice has been implemented on a given project, you are probably: (1) trying to exceed the necessary threshold to conclude that this is part of the “easy stuff” I mentioned above; (2) trying to bluff your way through a weak implementation by burying the appraisal team member in tangentially relevant stuff; or (3) misunderstanding the practice. Trust me, if the appraisal team does not believe they have sufficient data to reach the same conclusion you did â€" that the practice has been implemented on a given project â€" then they’ll be asking you for information that covers the gap â€" and THAT may be done in the interviews.
      >
      > The discussion above does touch on an important point â€" and that is that before you conduct the SCAMPI A, you should have convinced YOURSELVES that you have truly achieved the target maturity level. A successful outcome should be your expected result â€" after all, if YOU aren’t convinced that you DESERVE the maturity level then you’re really just gambling â€" rather than investing in a validation event.
      >
      > Conducting a preliminary appraisal is always a good idea as it (1) helps the lead appraiser and other external appraisal team members understand the organization and it’s jargon; (2) surfaces any model interpretation issues; (3) identifies any alternative practices and their relatively strength/weakness; and (4) identifies areas of probable concern with respect to a successful outcome. In addition, it should guide your future PIIDing activities as described above.
      >
      > Developing PIIDs was never intended to be a full time career. Use it as a means to convince yourselves that you have achieved the maturity level, and then share that reasoning with the appraisal team for validation. If you take that approach, things should work out just fine â€" for 75-90% anyway. And then we’ll have some interesting conversations about the rest...
      >
      > Regards,
      >
      > Pat
      >
      >
      >
      >
      > From: Ahmed Abd El Aziz
      > Sent: Tuesday, May 31, 2011 4:32 AM
      > To: cmmi_process_improvement@yahoogroups.com
      > Subject: RE: [CMMi Process Improvement] PI practices enlightenment
      >
      >
      >
      > Dear Greer,
      >
      >
      > It is my pleasure to meet you too.
      >
      >
      > Before being an LA I was an EPG Head in an ML3 organization and I suffered a lot with the PIID. In SCAMPI v1.3 a new approach is explained that saves a lot of effort which is called managed discovery. You may refer to the SCAMPI MDD document for details. Here I will briefly explain the approach I followed which I got Patrick O’Toole and Hillel Glazer in their participation in this group or in some conference. I can summarize it in the following points:
      >
      > 1. I usually conduct an early spot check â€" similar to SCAMPI C â€" with the organization. It is completely done in discovery mode. One of my deliverables to the company by the end of the Spot Check is an excel file containing mapping between CMMI practices and what I considered as a direct evidence for that practice. Actually it is part of the my workbook I use during the Spot Check.
      >
      > 1. When it comes to formal appraisal â€" SCAMPI A â€" I ask the company to fill the PIID with direct evidences only. This way we have spared almost half of the effort. The mapping file mentioned in the previous step is used gradually while projects implemented or just before the appraisal to fill the PIID. It safes a lot of time for them in thinking what are the appropriate evidence.
      >
      > 2. I ask the appraisal team to try to get 100% affirmation coverage.
      >
      > 3. For the practices we do not get affirmation, I ask the appraisal team to fill the coverage gaps with indirect evidences from the direct evidences they have already seen. Many evidences are considered as direct for some practices and indirect for others.
      >
      >
      > I also had another trial, but no one used it. I prepared excel file that contains a matrix where the documents (evidences) are in one dimension, and all the CMMI practices in the other dimension. EPG team has to fill the document names with the links only once. If the document is an evidence for some practice, this has to be mentioned in the interaction cell. As ATMs are not used to this view of data, I wrote a macro that converts the PIID from this simple data entry view to the traditional view (each PA has a separate sheet).
      >
      >
      > The templates I use are available and I can share it with you or anyone else or even with the whole group if people are interested. However I did not update them with the new SCAMPI v1.3 yet.
      >
      >
      > B. Regards
      >
      >
      > Ahmed Abd El Aziz
      >
      > Quality Expert
      >
      > PMP and SEI-Certified LA for DEV and SVC
      >
      > Software Engineering Competence Center "SECC"
      >
      > Information Technology Industry Development Agency "ITIDA"
      >
      >
      > Smart Village, Building (B121)
      >
      > Cairo-Alexandria Desert Road
      > Giza, Egypt 12577
      >
      > T +202 3534 5206/5207/5208
      >
      > F +202 3539 2134
      >
      > M +2010 177 8765
      >
      > aaziz@...
      >
      > www.secc.org.eg
      >
      >
      > From: cmmi_process_improvement@yahoogroups.com [mailto:cmmi_process_improvement@yahoogroups.com] On Behalf Of dagreergad
      > Sent: Monday, May 30, 2011 3:09 PM
      > To: cmmi_process_improvement@yahoogroups.com
      > Subject: Re: [CMMi Process Improvement] PI practices enlightenment
      >
      >
      >
      >
      > Greetings Ahmed, So you're an LA, pleasure to meet you.
      >
      > I've been apologizing to clients for years about the significant effort required to create PIIs, an artifact that seems to do the organization little good, except for making the appraisal easier. If you can come up with a better solution, please share it with the rest of us. Of course you could skip the PIIs all together and just conduct a discovery appraisal (permitted in all versions of the MDD).
      >
      > I estimate (WARNING - Made up statistic approaching) it would take nearly forever to conduct even a small appraisal if you had to hunt for all the direct evidence to review.
      >
      > Einstein said make everything as simple as possible and no simpler. It seems that you've done well by your clients to make PIIs as painless as possible, e.g. the mapping of what they do as broadly as possible to model components. It also seems that you've tried to reasonably reduce the amount of duplicative/redundant data required.
      >
      > The main challenge for a small organization is, and always has been, that they still have to present process and project evidence against a model that is identical in size and requirements to what Boeing has to provide, which necessarily means that it will seem proportionately more impacting.
      >
      > Maybe there could be a specific tailoring for small companies where they would only have to provide data at the goal level, that would reduce the data requirements. I don't think that would satisfy anyone, and nobody would do business with them, so that's probably not the way to go.
      >
      > I think that we should continue to look for ways to make PII development more efficient so that when a small company eventually needs to conduct an appraisal, it can be done more efficiently. That way still seems to be to make the PIIs part of a project's startup and continuing costs. That way the cost is absorbed over time, in little pieces.
      >
      > Just my thoughts.
      >
      > Best regards,
      >
      > DAGreer
      > mailto:dagreer%40asses-it.com
      >
      > --- In mailto:cmmi_process_improvement%40yahoogroups.com, Ahmed Abd El Aziz <aaziz@> wrote:
      > >
      > > Ed,
      > >
      > > That is exactly the point. I am on the other side. I am the Lead Appraiser who always reviews or does the mapping with the appraisal team and the EPG Team. This was always the case in all ML3 appraisals I conducted in small settings.
      > >
      > > When I asked for the possibility of making it NA, was not because it is actually not applicable, but because it adds overhead to the appraisal in small settings. No value added to the organization at all in such appraisals. It just shows how the EPG team can map work to CMMI practices of PI and how the appraisal team approves this mapping.
      > >
      > > B. Regards
      > >
      > > Ahmed Abd El Aziz
      > > Quality Expert
      > > PMP and SEI-Certified LA for DEV and SVC
      > > Software Engineering Competence Center "SECC"
      > > Information Technology Industry Development Agency "ITIDA"
      > >
      > > Smart Village, Building (B121)
      > > Cairo-Alexandria Desert Road
      > > Giza, Egypt 12577
      > > T +202 3534 5206/5207/5208
      > > F +202 3539 2134
      > > M +2010 177 8765
      > > aaziz@
      > > www.secc.org.eg<http://www.secc.org.eg/>
      > >
      > > From: mailto:cmmi_process_improvement%40yahoogroups.com [mailto:mailto:cmmi_process_improvement%40yahoogroups.com] On Behalf Of EDWARD F WELLER III
      > > Sent: Monday, May 30, 2011 4:20 AM
      > > To: mailto:cmmi_process_improvement%40yahoogroups.com
      > > Subject: Re: [CMMi Process Improvement] PI practices enlightenment
      > >
      > >
      > > Ahmed
      > > Then you are doing PI - start from what you do and map to the model, not the reverse. If you are doing the activities as part of design (the work), and satisfy something in PI (the model), does anyone care other than the appraisal team ? Below the process mapping done by the EPG, I have always tried to minimize knowledge of the CMMI as the CMMI with the people doing the work
      > >
      > > Ed
      > > ----- Original Message -----
      > > From: Ahmed Abd El Aziz<mailto:aaziz@>
      > > To: mailto:cmmi_process_improvement%40yahoogroups.com<mailto:mailto:cmmi_process_improvement%40yahoogroups.com>
      > > Sent: Saturday, May 28, 2011 11:08 AM
      > > Subject: RE: [CMMi Process Improvement] PI practices enlightenment
      > >
      > >
      > > Dear Ed,
      > > I did not mean there is not interfaces. Of course there are, but they are handled in a simple form by other PAs. For example human user interface are usually identified in the requirements and implemented in the design. In most of the cases there is a small development team (no more than three developers) and the development system is almost one component. There is integration sequence other than the MAK file or similar. There is no special environment for integration that has a stub or something similar that is needed only at the integration time and not needed later. The development environment is enough.
      > > When I read PI in light of large scale systems (many teams and many technical components) and software engineering, this PA pays my attention to a lot of important practices that cannot be handled via traditional requirements and development activities. On the other hand, when I look to small systems (one small team and group of classes, not components), nothing special to be done to implement the PI. It becomes the appraisal team's job to map the organizational practices to the PA practices.
      > > B. Regards
      > > Ahmed Abd El Aziz
      > > Quality Expert
      > > PMP and SEI-Certified LA for DEV and SVC
      > > Software Engineering Competence Center "SECC"
      > > Information Technology Industry Development Agency "ITIDA"
      > > Smart Village, Building (B121)
      > > Cairo-Alexandria Desert Road
      > > Giza, Egypt 12577
      > > T +202 3534 5206/5207/5208
      > > F +202 3539 2134
      > > M +2010 177 8765
      > > aaziz@
      > > www.secc.org.eg<http://www.secc.org.eg/>
      > > From: mailto:cmmi_process_improvement%40yahoogroups.com [mailto:mailto:cmmi_process_improvement%40yahoogroups.com] On Behalf Of EDWARD F WELLER III
      > > Sent: Saturday, May 28, 2011 6:20 PM
      > > To: mailto:cmmi_process_improvement%40yahoogroups.com
      > > Subject: Re: [CMMi Process Improvement] PI practices enlightenment
      > >
      > > Ahmed
      > > If you can demonstrate work that has no interfaces of any kind for the work you are doing, then PI might be not applicable
      > > But then you might have a product no one would buy as it would have no interface to the buyer !!!
      > > Read Pat's reply carefully. Even the most simple of work (or maybe the most difficult), one line fixes to a failing app ot program, have interfaces and an integration sequence (else how does it get to the customer)
      > > Ed
      > > ----- Original Message -----
      > > From: Ahmed Abd El Aziz<mailto:aaziz@>
      > > To: mailto:cmmi_process_improvement%40yahoogroups.com<mailto:mailto:cmmi_process_improvement%40yahoogroups.com>
      > > Sent: Saturday, May 28, 2011 8:57 AM
      > > Subject: RE: [CMMi Process Improvement] PI practices enlightenment
      > >
      > > Dear Juan,
      > > I share you your dream. In many ML3 appraisals I conducted where the project and the team are small, we had a hard time to accept evidences for PI practices even much simpler than the evidences you provided for most of the PI practices.
      > > To modify the dream a little, I hope SEI permits PI to be Not Applicable - same like SAM.
      > > B. Regards
      > > Ahmed Abd El Aziz
      > > Quality Expert
      > > PMP and SEI-Certified LA for DEV and SVC
      > > Software Engineering Competence Center "SECC"
      > > Information Technology Industry Development Agency "ITIDA"
      > > Smart Village, Building (B121)
      > > Cairo-Alexandria Desert Road
      > > Giza, Egypt 12577
      > > T +202 3534 5206/5207/5208
      > > F +202 3539 2134
      > > M +2010 177 8765
      > > aaziz@
      > > www.secc.org.eg<http://www.secc.org.eg/>
      > > From: mailto:cmmi_process_improvement%40yahoogroups.com [mailto:mailto:cmmi_process_improvement%40yahoogroups.com] On Behalf Of jctorres.geo
      > > Sent: Friday, May 27, 2011 7:25 PM
      > > To: mailto:cmmi_process_improvement%40yahoogroups.com
      > > Subject: [CMMi Process Improvement] PI practices enlightenment
      > >
      > >
      > > Hello everyone, greetings from Peru!!
      > >
      > > I have a dream, and I want to share it with all of you (tears), that the SEI remove the PI process area.... :) My therapist will have an easier time.
      > >
      > > Sorry for the joke, but it reflects some frustration I have with PI practices.
      > >
      > > In fact, my problems have narrowed to a couple of PI practices, and I will appreciate your comments and suggestions on how to implement them:
      > >
      > > Some background info: We code in .NET C#. The internal interfaces we consider critical are those defined by classes that provide services (methods) to other classes ("interface classes"). Let me briefly explain how we are implementing the following PI practices:
      > >
      > > 1) SP 2.1 Review Interface Descriptions for Completeness
      > > We use a UML class diagram that shows the interface classes and their methods, along with other type of classes. It shows the relationships between them.
      > >
      > > This diagram is prepared using a modelling tool and established as a starting point before programming. It is reviewed by the team at the beginning of the development stage.
      > >
      > > This diagram is maintained because of 2 reasons:
      > > - Changes in functionality or technical solution require changes at class levels and interfaces, so the diagram gets updated.
      > > - New classes or methods may be inserted by the programmers directly into the source code (mainly because of programming decisions). An analyst, periodically, will execute reverse-engineering from the source code to the diagram (using the modelling tool), to update the diagram.
      > >
      > > Those are the activities that we consider to fit into this practice.
      > >
      > > 2) SP 2.2 Manage Interfaces
      > > We consider that internal interfaces are managed because of the fact that the interface classes diagram is managed by a modelling tool (which is stored in a repository with check in/out features), and that it is updated by the steps explained before.
      > >
      > > Additionally, the methods of the classes are documented in the source code. The documentation include the goal of the interface, the description of the parameters and output.
      > >
      > > Thank you very much for your help
      > >
      > > Juan Carlos
      > >
      >
    • Raman Raman
      Thanks to the various forum members who responded and shared their perspectives. Appreciated. Ram ... From: Jeff Dalton Subject:
      Message 42 of 42 , Jun 22, 2011
      • 0 Attachment
        Thanks to the various forum members who responded and shared their perspectives.

        Appreciated.

        Ram


        --- On Mon, 6/20/11, Jeff Dalton <jeff@...> wrote:

        From: Jeff Dalton <jeff@...>
        Subject: Re: [CMMi Process Improvement] ICMMI training on Videoconference
        To: cmmi_process_improvement@yahoogroups.com
        Date: Monday, June 20, 2011, 12:01 AM

         

        Perhaps there are ways to deliver CMMI training over the web, but this class is not one of them.  When I teach this class I spend most of the time NOT talking about the slides.  Interaction and real-time examples resonate the most with students and that's why they are there.  A web delivery just doesn't cut it.

        If you want to become familiar enough with the CMMI to really get started, a F2F class with a good instructor, who has a lot of practical experience, and an engaged class, in an environment that promotes learning, is the minimum price of admission.



        Jeff Dalton
        President, Broadsword

        Certified SCAMPI Lead Appraiser
        Certified CMMI Instructor
        Certified CMMI for Practitioners Instructor
        Scrum Master

        Office: 248.341.3367
        Fax:      248.341.3672
        `

        Read Jeff's writings at http://askTheCMMIAppraiser.com

        Follow Jeff's Tweets: CMMIAppraiser



        From: Heather Oppenheimer <hoppenheimer@...>
        Reply-To: <cmmi_process_improvement@yahoogroups.com>
        Date: Sun, 19 Jun 2011 10:59:52 -0400
        To: <cmmi_process_improvement@yahoogroups.com>
        Subject: RE: [CMMi Process Improvement] ICMMI training on Videoconference

         

        Joe, that’s certainly a risk with distance learning options.   There are several other risks and issues that would have to be addressed as well.    However, I think there are legitimate reasons to ask about distance learning options, especially for people who would need to travel to a course, rather than having it delivered at their own location.  Not only do they have the costs related to travel, but the course then requires additional time (getting there/home, evenings away from family and other personal commitments, etc.)

         

        However, I think the main reason the courses – both Intro-DEV and Intro-SVC – are required to be delivered face to face – is because they were *designed and developed* as face-to-face, instructor-led courses.  It would require significant development work to redesign and redevelop the courses so that they would provide value in a distance learning environment.  As far as I know, that work isn’t budgeted or planned at this time.

         

        Yes, if an instructor does nothing more than read PowerPoint slides to the class, it would work just as well to do that in a webinar setting (and I wouldn’t blame people for doing other things at the same time, if that were the case!)  However, the course includes a lot of interaction among the participants as well as between the instructor and the participants.  Most of the in-progress feedback that an instructor relies on and the group exercises are based on face to face interaction, and wouldn’t work in a video-teleconference.

         

        Unless the work to redesign and redevelop the 3 day Intro courses becomes a priority and has a budget, they will remain “face to face” only.

         

        Cheers

        Heather Oppenheimer

        www.oppenpartners.com

         

         

         

        From: cmmi_process_improvement@yahoogroups.com [mailto:cmmi_process_improvement@yahoogroups.com] On Behalf Of Joe Smith
        Sent: Sunday, June 19, 2011 6:35 AM
        To: cmmi_process_improvement@yahoogroups.com
        Subject: Re: [CMMi Process Improvement] ICMMI training on Videoconference

         

         

         

        I have heard others in our org ask this too. Why does it seem as if people want to take this class in a way they probably will not anything learn from, and take the class while they are doing thier "real" work. Is this just to get credit while being able to ignore it even more? We have folks who take other training all the time like this, heck they attend all sorts of meetings like this  - they are on the phone or do thier email the whole time. When you ask them a question the answer is always like - oh I'm sorry, I just missed what you said, can you repeat it?

        --- On Fri, 6/17/11, Raman Raman <raman_rr@...> wrote:

        From: Raman Raman <raman_rr@...>
        Subject: [CMMi Process Improvement] ICMMI training on Videoconference
        To: cmmi_process_improvement@yahoogroups.com
        Date: Friday, June 17, 2011, 4:46 PM


        Is it possible to have the SEI Authorised Introduction to CMMI training conducted on videoconference or telepresence or an equivalent mechanism ?

        Please suggest. Thanks

        Ram

      Your message has been successfully submitted and would be delivered to recipients shortly.