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

Re: [XP] Digest Number 3653

Expand Messages
  • aacockburn
    ... organization. That just means making sure that people know what to do, how to do it, and when. ... them XP and about the management of the projects? ...
    Message 1 of 9 , Sep 30, 2003
    • 0 Attachment
      --- In extremeprogramming@yahoogroups.com, "Hillel Glazer, Entinex,
      The Technology Strategy Company" <online@e...> wrote:
      > CMM/CMMI is big on "institutionalizing" the processes within an
      organization. That just means making sure that people know what to
      do, how to do it, and when.
      > But in brass tacks... if I were appraising an XP shop, here's how
      it would go down:
      > - "if a new hire comes in that's never done XP, how do you teach
      them XP and about the management of the projects?"
      > - "what materials or approach do you use to train the new hires?"
      >
      > in CMMI, the separation in monitoring needs only be as "far" as
      someone who the PM doesn't hold administrative or technical sway,...

      Let's see if I can recap this in words I am more likely to use later:
      --->For an XP group to pass CMMI level 3 certification, they need
      ---> to show that someone NOT in the direct PM control chain
      ---> is monitoring whether or not the group is doing XP,
      ---> AND the monitoring person has to return a positive result.

      I claim that
      (1) There are no XP projects in the world where someone
      outside the direct PM control chain is actively monitoring
      whether the group is doing XP, and
      (2) No monitor anywhere in the world knows if any group is doing
      XP anyway (because there are no accepted compliance criteria).


      Alistair
    • yahoogroups@jhrothjr.com
      ... From: aacockburn To: extremeprogramming@yahoogroups.com
      Message 2 of 9 , Sep 30, 2003
      • 0 Attachment
        ----- Original Message -----
        From: "aacockburn" <acockburn.at.aol.com@...>
        To: "extremeprogramming@yahoogroups.com"
        <extremeprogramming.at.yahoogroups.com@...>
        Sent: Tuesday, September 30, 2003 3:17 AM
        Subject: Re: [XP] Digest Number 3653


        > --- In extremeprogramming@yahoogroups.com, "Hillel Glazer, Entinex,
        > The Technology Strategy Company" <online@e...> wrote:
        > > CMM/CMMI is big on "institutionalizing" the processes within an
        > organization. That just means making sure that people know what to
        > do, how to do it, and when.
        > > But in brass tacks... if I were appraising an XP shop, here's how
        > it would go down:
        > > - "if a new hire comes in that's never done XP, how do you teach
        > them XP and about the management of the projects?"
        > > - "what materials or approach do you use to train the new hires?"
        > >
        > > in CMMI, the separation in monitoring needs only be as "far" as
        > someone who the PM doesn't hold administrative or technical sway,...
        >
        > Let's see if I can recap this in words I am more likely to use later:
        > --->For an XP group to pass CMMI level 3 certification, they need
        > ---> to show that someone NOT in the direct PM control chain
        > ---> is monitoring whether or not the group is doing XP,
        > ---> AND the monitoring person has to return a positive result.
        >
        > I claim that
        > (1) There are no XP projects in the world where someone
        > outside the direct PM control chain is actively monitoring
        > whether the group is doing XP, and
        > (2) No monitor anywhere in the world knows if any group is doing
        > XP anyway (because there are no accepted compliance criteria).

        I seem to remember a paper by Kent Beck with the rather interesting
        title: "Are you being XPed On?" that set out a set of criteria that he could
        use simply by walking into the team's room.

        I'm still in the process of digesting "Lean Software Development."
        My take on the question of whether you're doing XP or not is that
        XP is distinguished from other Agile methodologies by the software
        engineering discipline in use, and that discipline has one major
        characteristic:

        A single developer (with, of course, a pair) takes a story
        (which does not usually exceed three pair-days) from requirements
        to fully integrated, deployable code in one step, without any handoffs.

        Everything else in the methodology is in support of this.

        So I've really got two answers to your second point. One is
        that my criteria for determining if a team is using XP depends
        on whether that one thing is being done well: that is, stories
        are being implemented with minimal defects and I can walk
        in at any time and pick up a deployable install with at most n
        (where n is the # of developers) partially implemented stories.

        As it turns out, XP is the only methodology that I know of
        where I can do this. I can't with SCRUM because there's
        no requirement that they have a deployable at any point
        other than the end of the sprint. I believe that's also true
        of Crystal.

        The other is that CMMi (and CMM) don't care whether
        you're using a name brand methodology or not. If the auditor
        is doing his job, the only effect of someone saying "I'm using
        XP" (or "I'm using fubar", or whatever) is that he can pull a
        list of things out of his back pocket that are known trouble
        spots with that methodology. Otherwise, it simply shouldn't
        matter.

        John Roth

        >
        >
        > Alistair
        >
        >
        >
        >
        >
        >
        >
        >
        > To Post a message, send it to: extremeprogramming@...
        >
        > To Unsubscribe, send a blank message to:
        extremeprogramming-unsubscribe@...
        >
        > ad-free courtesy of objectmentor.com
        >
        > Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/
        >
        >
      • Ron Jeffries
        ... No. One does not have to be doing XP to be CMM certified. One has to be emitting certain evidence or artifacts addressing the CMM KPAs. In the paragraph
        Message 3 of 9 , Sep 30, 2003
        • 0 Attachment
          On Tuesday, September 30, 2003, at 3:17:46 AM, aacockburn wrote:

          > Let's see if I can recap this in words I am more likely to use later:
          > --->For an XP group to pass CMMI level 3 certification, they need
          > ---> to show that someone NOT in the direct PM control chain
          > ---> is monitoring whether or not the group is doing XP,
          > ---> AND the monitoring person has to return a positive result.

          No. One does not have to be doing XP to be CMM certified. One has to be
          emitting certain evidence or artifacts addressing the CMM KPAs.

          In the paragraph below you are claiming that no XP project in the world is
          CMM L3. It seems impossible that one could know that.

          > I claim that
          > (1) There are no XP projects in the world where someone
          > outside the direct PM control chain is actively monitoring
          > whether the group is doing XP, and

          The individual would merely have to be actively monitoring whether they are
          producing the required evidence, not whether they were doing XP, which
          makes the following moot.

          > (2) No monitor anywhere in the world knows if any group is doing
          > XP anyway (because there are no accepted compliance criteria).

          This is perhaps why there is no XP project certification entity. The
          remarks above do not address whether an XP team could be CMM L3.

          They might, however, address whether there is such a thing as an "XP team",
          but both Beck's and my definition of XP do seem to allow for that
          possibility.

          Ron Jeffries
          www.XProgramming.com
          Inigo Montoya: You are wonderful!
          Man in Black: Thank you. I have worked hard to become so.
        • Ron Jeffries
          ... This is a very interesting characterization. I would say that it was certainly not implicit in Greco-Roman XP, and I m not aware of it being called out
          Message 4 of 9 , Sep 30, 2003
          • 0 Attachment
            On Tuesday, September 30, 2003, at 7:10:15 AM, yahoogroups@... wrote:

            > Alistair wrote:
            >> (2) No monitor anywhere in the world knows if any group is doing
            >> XP anyway (because there are no accepted compliance criteria).

            > ...

            > I'm still in the process of digesting "Lean Software Development."
            > My take on the question of whether you're doing XP or not is that
            > XP is distinguished from other Agile methodologies by the software
            > engineering discipline in use, and that discipline has one major
            > characteristic:

            > A single developer (with, of course, a pair) takes a story
            > (which does not usually exceed three pair-days) from requirements
            > to fully integrated, deployable code in one step, without any handoffs.

            This is a very interesting characterization. I would say that it was
            certainly not implicit in "Greco-Roman" XP, and I'm not aware of it being
            called out in any of the classical or modern printed literature, but it
            does go to an important, arguably key aspect of XP.

            > ...

            > As it turns out, XP is the only methodology that I know of
            > where I can do this. I can't with SCRUM because there's
            > no requirement that they have a deployable at any point
            > other than the end of the sprint. I believe that's also true
            > of Crystal.

            Yes. Of course a team doing C or S could do those things anyway. Some teams
            using Scrum on the outside use XP on the inside, e.g. XP@Scrum. (I don't
            know why they don't just use XP planning as their interface, but oh well.)

            > The other is that CMMi (and CMM) don't care whether
            > you're using a name brand methodology or not. If the auditor
            > is doing his job, the only effect of someone saying "I'm using
            > XP" (or "I'm using fubar", or whatever) is that he can pull a
            > list of things out of his back pocket that are known trouble
            > spots with that methodology. Otherwise, it simply shouldn't
            > matter.

            Yes, exactly. CMM doesn't verify whether you're doing XP, it verifies
            whether you are meeting the KPAs. The auditor, upon hearing that you're
            doing XP, would ask various questions about the KPAs and how they are met.
            Any team, XP or not, desiring to be assessed at some level, has to have
            answers to all the KPAs.

            Getting those answers ready is a fair amount of work, but well within human
            reach. A good preparation for CMM (I haven't brought myself up to speed on
            CMMi) can be done by going through the KPAs and thinking how one will
            answer. Here's an extended example of that thinking:

            ===============================
            The auditor may ask:

            At Level 2, KPA 7.1, Requirements Management asks that you create a
            common understanding between the customer and the software project of the
            project's requirements. Commitment 1 of that KPA asks that your project
            follow a written organizational policy for managing the system
            requirements. May I see that policy?

            And the team hands the auditor a 3x5 card that says "We manage requirements
            using the XP Planning Game," saying "That's our policy document." The
            auditor would likely ask for more detail and be directed to the pink book,
            the green book, a training video, or a company web page describing the
            practice. The training video would be coolest, as it would essentially
            propagate an oral tradition in a form that the CMM would surely accept.

            The auditor would know that the management of requirements "typically"
            specifies that they are documented (in XP on cards, ancillary documents as
            required, and written acceptance tests), and that they are reviewed by the
            software managers and other affected groups. The XP Customer [Role] would
            need to include the "other affected groups", and one might require that the
            manager attend the planning meetings, or look at the project whiteboard.

            The KPA also requires that the "software plans, work products, and
            activities are changed to be consistent with changes to the allocated
            requirements." XP does this automatically, as requirements only come in
            three states: Done, In Process In Current Iteration, and Unallocated. Done
            requirements never change; Unallocated requirements can change all they
            want because no one is impacted; In Process requirements come already with
            the necessary documentation, per the paragraph above.

            Moving on to L3, the Auditor might ask:

            KPA 8.1 requires a group that is responsible for the organization's
            software process activities. What is your group?

            And the company (note that almost none of the L3 KPAs are team-related)
            replies:

            We are aware that the CMM states:

            A group is the collection of departments, managers, and individuals who
            have responsibiliyt for a set of tasks or activities. A group could
            vary from a single individual assigned part time, to several part-time
            individuals assigned from different departments, to several individuals
            dedicated full time.

            We know that the CMM states that where possible, this group is staffed by
            a full-time core, and that it is capable of addressing requirements
            analysis, software design, coding, testing, configuration management and
            QA disciplines.

            In our company, we call this our Software Engineering Process Group, and
            it consists of volunteers from the development departments and is
            formally chaired by the CTO. Current members are Joe, Lucinda, Candace,
            and David.

            And as far as I can tell, the auditor would say either "Oh," or "Cool," and
            go on to ask about funding for the group, which would be answered in terms
            of time members could and would spend; and about training, which is a
            separate KPA.
            ==================================

            I see no evidence that it is impossible for a team doing XP to be assessed
            at L3. One just has to do certain things. XP covers some of them more or
            less automatically, and one would add the necessary practices to do the
            rest. I don't see why it would be a problem at all.

            Ron Jeffries
            www.XProgramming.com
            Ron Jeffries, speaking for Boskone ... Out.
          • yahoogroups@jhrothjr.com
            ... From: Ron Jeffries To: extremeprogramming@yahoogroups.com
            Message 5 of 9 , Sep 30, 2003
            • 0 Attachment
              ----- Original Message -----
              From: "Ron Jeffries"
              <ronjeffries.at.XProgramming.com@...>
              To: "extremeprogramming@yahoogroups.com"
              <extremeprogramming.at.yahoogroups.com@...>
              Sent: Tuesday, September 30, 2003 8:54 AM
              Subject: Re: [XP] Digest Number 3653


              > On Tuesday, September 30, 2003, at 7:10:15 AM, yahoogroups@...
              wrote:
              >
              > > Alistair wrote:
              > >> (2) No monitor anywhere in the world knows if any group is doing
              > >> XP anyway (because there are no accepted compliance criteria).
              >
              > > ...
              >
              > > I'm still in the process of digesting "Lean Software Development."
              > > My take on the question of whether you're doing XP or not is that
              > > XP is distinguished from other Agile methodologies by the software
              > > engineering discipline in use, and that discipline has one major
              > > characteristic:
              >
              > > A single developer (with, of course, a pair) takes a story
              > > (which does not usually exceed three pair-days) from requirements
              > > to fully integrated, deployable code in one step, without any handoffs.
              >
              > This is a very interesting characterization. I would say that it was
              > certainly not implicit in "Greco-Roman" XP, and I'm not aware of it being
              > called out in any of the classical or modern printed literature, but it
              > does go to an important, arguably key aspect of XP.


              It took me quite a while to realize that it was the key element.
              Part of that was reading "Lean Software Development," which
              I mentioned, and part of it was the "Herding Racehorses,
              Racing Sheep" presentation that Dave Thomas gave us at the
              Atlanta Spin meeting (or maybe the Java User's Group the next
              day) about the various levels of skills acquisition.

              One of the key realizations was Mary Poppendicks's comment
              in "Lean Software Development" to the effect that "Fixed Length
              Iterations is the starting point that always works" because it forces
              a complete and pervasive change in the way you do everything
              else. So I looked through the XP practices for what the core
              element was (or elements were) that forced the rest of the
              practices. The only one I could find that had that kind of
              pervasive effect was organizing the effort as a pull system:
              an availible developer takes the next work item off the top
              of the stack, works on it until it is done, and then takes the
              next item, while reducing or eliminating the Waste inherent
              in handoffs and having intermediate work products in
              inventory.

              >
              > > ...
              >
              > > As it turns out, XP is the only methodology that I know of
              > > where I can do this. I can't with SCRUM because there's
              > > no requirement that they have a deployable at any point
              > > other than the end of the sprint. I believe that's also true
              > > of Crystal.
              >
              > Yes. Of course a team doing C or S could do those things anyway. Some
              teams
              > using Scrum on the outside use XP on the inside, e.g. XP@Scrum. (I don't
              > know why they don't just use XP planning as their interface, but oh well.)

              Well, that's true, but you can't say that SCRUM does this because
              it doesn't *require* it. I can say I'm doing SCRUM, and you can't
              make any a priori judgement about what I'm doing inside a sprint.

              > > The other is that CMMi (and CMM) don't care whether
              > > you're using a name brand methodology or not. If the auditor
              > > is doing his job, the only effect of someone saying "I'm using
              > > XP" (or "I'm using fubar", or whatever) is that he can pull a
              > > list of things out of his back pocket that are known trouble
              > > spots with that methodology. Otherwise, it simply shouldn't
              > > matter.
              >
              > Yes, exactly. CMM doesn't verify whether you're doing XP, it verifies
              > whether you are meeting the KPAs. The auditor, upon hearing that you're
              > doing XP, would ask various questions about the KPAs and how they are met.
              > Any team, XP or not, desiring to be assessed at some level, has to have
              > answers to all the KPAs.
              >
              > Getting those answers ready is a fair amount of work, but well within
              human
              > reach. A good preparation for CMM (I haven't brought myself up to speed on
              > CMMi) can be done by going through the KPAs and thinking how one will
              > answer.

              Interestingly, it could be a very good exercise even if one doesn't ever
              intend to get formal certification. I'm reminded of a management article
              from the '60s that recommended doing a system study for replacing a
              manual system with an automated one, and then cancelling the computer!
              The reason was that the study itself would shake enough improvement
              items out that you didn't need to go through the pain of automating.

              Here's an extended example of that thinking:

              [snip a really excellent example!]

              > I see no evidence that it is impossible for a team doing XP to be assessed
              > at L3. One just has to do certain things. XP covers some of them more or
              > less automatically, and one would add the necessary practices to do the
              > rest. I don't see why it would be a problem at all.

              And some of them would probably lead to really useful process improvement.
              Some wouldn't, of course.

              > Ron Jeffries
              > www.XProgramming.com
              > Ron Jeffries, speaking for Boskone ... Out.

              Xa ^H^H John Roth speaking for Ploor... Out.
            • aacockburn
              ... he could ... Anyone know where this article is now? I couldn t find it just now. It would be nice if those criteria could be used by other people. Your
              Message 6 of 9 , Sep 30, 2003
              • 0 Attachment
                --- In extremeprogramming@yahoogroups.com, yahoogroups@j... wrote:
                --- In extremeprogramming@yahoogroups.com, yahoogroups@j... wrote:
                > I seem to remember a paper by Kent Beck with the rather interesting
                > title: "Are you being XPed On?" that set out a set of criteria that
                he could
                > use simply by walking into the team's room.

                Anyone know where this article is now? I couldn't find it just now.
                It would be nice if those criteria could be used by other people.


                Your much more important point, which I had completely overlooked, is
                this:

                > CMMi (and CMM) don't care whether
                > you're using a name brand methodology or not. If the auditor
                > is doing his job, the only effect of someone saying "I'm using
                > XP" (or "I'm using fubar", or whatever) is that he can pull a
                > list of things out of his back pocket that are known trouble
                > spots with that methodology. Otherwise, it simply shouldn't
                > matter.

                We ask whether XP can be assessed at CMMi level 3, but the correct
                question, which keeps slipping my mind, is that assessment is done of
                the team, not the methodology description.

                Therefore, the team wouldn't say to the assessor: "We're doing XP."
                They would say, "We're doing xxxx and yyyyy and zzzz." and by
                coincidence those might (or might not) be canonical XP. The assessor
                will only check whether they do what they say, in the CMMi checking
                format.

                So you are correct, and my point #2 (nobody knows what XP is) goes
                out the window.

                It still leaves point #1, which is that any team doing some variant
                on XP will still have to have the external person in place monitoring
                that they do what they say they do.

                Finally, I love this:

                > A single developer (with, of course, a pair) takes a story
                > (which does not usually exceed three pair-days) from requirements
                > to fully integrated, deployable code in one step, without any
                > handoffs.


                Thanks - Alistair
              • Robert Blum
                Hi Alistair! ... http://groups.yahoo.com/group/extremeprogramming/files/Are%20You%20Getting%20XPd%20On.pdf - Robert -- NEU FÜR ALLE - GMX MediaCenter - für
                Message 7 of 9 , Sep 30, 2003
                • 0 Attachment
                  Hi Alistair!

                  > > I seem to remember a paper by Kent Beck with the rather interesting
                  > > title: "Are you being XPed On?" that set out a set of criteria that
                  > he could
                  > > use simply by walking into the team's room.
                  >
                  > Anyone know where this article is now? I couldn't find it just now.
                  > It would be nice if those criteria could be used by other people.

                  It's right here:
                  :)

                  http://groups.yahoo.com/group/extremeprogramming/files/Are%20You%20Getting%20XPd%20On.pdf

                  - Robert

                  --
                  NEU FÜR ALLE - GMX MediaCenter - für Fotos, Musik, Dateien...
                  Fotoalbum, File Sharing, MMS, Multimedia-Gruß, GMX FotoService

                  Jetzt kostenlos anmelden unter http://www.gmx.net

                  +++ GMX - die erste Adresse für Mail, Message, More! +++
                • yahoogroups@jhrothjr.com
                  ... From: aacockburn To: extremeprogramming@yahoogroups.com
                  Message 8 of 9 , Sep 30, 2003
                  • 0 Attachment
                    ----- Original Message -----
                    From: "aacockburn" <acockburn.at.aol.com@...>
                    To: "extremeprogramming@yahoogroups.com"
                    <extremeprogramming.at.yahoogroups.com@...>
                    Sent: Tuesday, September 30, 2003 8:05 PM
                    Subject: Re: [XP] Digest Number 3653


                    > --- In extremeprogramming@yahoogroups.com, yahoogroups@j... wrote:
                    > --- In extremeprogramming@yahoogroups.com, yahoogroups@j... wrote:
                    > > I seem to remember a paper by Kent Beck with the rather interesting
                    > > title: "Are you being XPed On?" that set out a set of criteria that
                    > he could
                    > > use simply by walking into the team's room.

                    > Anyone know where this article is now? I couldn't find it just now.
                    > It would be nice if those criteria could be used by other people.

                    It's in the files section of this group.

                    > Your much more important point, which I had completely overlooked, is
                    > this:
                    >
                    > > CMMi (and CMM) don't care whether
                    > > you're using a name brand methodology or not. If the auditor
                    > > is doing his job, the only effect of someone saying "I'm using
                    > > XP" (or "I'm using fubar", or whatever) is that he can pull a
                    > > list of things out of his back pocket that are known trouble
                    > > spots with that methodology. Otherwise, it simply shouldn't
                    > > matter.
                    >
                    > We ask whether XP can be assessed at CMMi level 3, but the correct
                    > question, which keeps slipping my mind, is that assessment is done of
                    > the team, not the methodology description.
                    >
                    > Therefore, the team wouldn't say to the assessor: "We're doing XP."
                    > They would say, "We're doing xxxx and yyyyy and zzzz." and by
                    > coincidence those might (or might not) be canonical XP. The assessor
                    > will only check whether they do what they say, in the CMMi checking
                    > format.
                    >
                    > So you are correct, and my point #2 (nobody knows what XP is) goes
                    > out the window.
                    >
                    > It still leaves point #1, which is that any team doing some variant
                    > on XP will still have to have the external person in place monitoring
                    > that they do what they say they do.

                    I think that goes to the heart of organizational institutionalization, which
                    I belive is one of the issues in CMMi.

                    When I write software, I'm going to expend a significant amount
                    of effort writing and revising tests, and I'm going to expend a
                    significant proportion of my hardware resources running and
                    rerunning those tests. The monitoring function is simply tests
                    on an organization level to insure that the procedure remains
                    in compliance with the CMM/CMMi.

                    We don't expect to be able to write working software without
                    tests, we don't expect to manufacture saleable widgits without
                    tests, why should we expect to maintain a functioning organization
                    without tests? Ask an accountant about that one!

                    > Finally, I love this:
                    >
                    > > A single developer (with, of course, a pair) takes a story
                    > > (which does not usually exceed three pair-days) from requirements
                    > > to fully integrated, deployable code in one step, without any
                    > > handoffs.

                    It took me quite a while to identify that as one of the key
                    elements.

                    Thank you!
                    John Roth

                    >
                    > Thanks - Alistair
                    >
                  Your message has been successfully submitted and would be delivered to recipients shortly.