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

Iterations harmful?

Expand Messages
  • aacockburn
    At the Northern Utah HCI meeting last night I got to perform a bit of my rant against iterations (as opposed to deliveries ). There seemed to be some
    Message 1 of 4 , Sep 9, 2005
    • 0 Attachment
      At the Northern Utah HCI meeting last night I got to perform a bit of
      my rant against "iterations" (as opposed to "deliveries"). There seemed
      to be some sympathy pains in the audience, as well as some discussion.

      I wrote it up as an article, and posted it
      and the URL itself is
      http://alistair.cockburn.us/crystal/articles/aih/areiterationshazardous.
      htm

      If that URL wraps, as I expect it will, it'll be the top article at
      http://alistair.cockburn.us/sitehistory.htm

      happy reading --- Alistair
    • Ron Jeffries
      ... Very nice paper, and it is well-described as a rant. I myself, as all here know, eschew controversy, so I might have gone a different way. ;- I
      Message 2 of 4 , Sep 10, 2005
      • 0 Attachment
        On Saturday, September 10, 2005, at 1:42:22 AM, aacockburn wrote:

        > At the Northern Utah HCI meeting last night I got to perform a bit of
        > my rant against "iterations" (as opposed to "deliveries"). There seemed
        > to be some sympathy pains in the audience, as well as some discussion.

        > I wrote it up as an article, and posted it
        > and the URL itself is
        > http://alistair.cockburn.us/crystal/articles/aih/areiterationshazardous.
        > htm

        > If that URL wraps, as I expect it will, it'll be the top article at
        > http://alistair.cockburn.us/sitehistory.htm

        Very nice paper, and it is well-described as a rant. I myself, as
        all here know, eschew controversy, so I might have gone a different
        way. ;-> I particularly like the word "deliveries": vide infra.

        Interesting ... I might suggest references to Y.T.'s articles on ...
        "running, tested, features", at
        http://www.xprogramming.com/xpmag/AgileTopDown.htm and
        http://www.xprogramming.com/xpmag/jatRtsMetric.htm

        The latter may be the first "published" occurrence of the RTF term.
        Not that I'm looking for primacy: the point is to get the idea out
        there. I'm happy to see that we agree on its importance.

        Second, the article states that no self-respecting agilist will
        permit iterations longer than two weeks. I think that at least some
        of the Scrum people still have some remaining shreds of
        self-respect, and the standard in Scrum is still a month. It's true
        that some Scrum folks such as Mike Cohn are now recommending two
        weeks. Since he could lift me over his head, I'm not going to argue
        self respect with Mike.

        Third, the article suggests that short iterations are a matter of
        pride. That is not the case in my opinion. AYMR, the original XP
        project started with three-week iterations, tried one week, and went
        back to three, having found it better. The /XP Installed/ book, I
        believe, recommends three weeks. (ICBW).

        The one-week XP iteration was adopted as a good idea by the XP
        leadership, not because ours are bigger than yours, though of course
        that goes without saying, but because XP teams all over the place
        were using one week iterations to good effect. The short iterations
        were based on experience. One might say that we got in front of the
        parade.

        One of the interesting effects I've observed relates to this. I
        encounter clients who, over the course of a month, have trouble
        completing features. I tried recommending that they go to one week
        iterations (I like the term "deliveries" very much, by the way, and
        will start using it as soon as I step out of the thought here.)

        An interesting thing happens when a team shifts to a one-week
        iteration, with intention to deliver RTF: they have to focus. When
        there are only five days till lift-off, it's easier to keep your eye
        on the ball. I think that it's harder for Parkinson's Law to take
        effect inside the shorter iteration. Whatever the cause, going to a
        shorter iteration seems to help these teams learn to be DONE.

        Some teams find the pace too maddening: C3 did, and some of my
        current client teams have gone back to two week deliveries. That's
        absolutely cool with me. To me, what matters is helping teams learn
        to see and adjust the controls that help them do better projects. I
        don't win when they use one-week cycles: I win when they're happy
        and productive.

        Neat article, good rant, good points. Thanks!

        Ron Jeffries
        www.XProgramming.com
        If you don't have the courage to say what you think,
        there isn't much use in thinking it, is there?
        --Thomas Jay Peckish II
      • aacockburn
        ... My bad, thanks for the catch, it s already fixed. I had that reference in the Governance paper; didn t detect its absence here. I care about these things,
        Message 3 of 4 , Sep 10, 2005
        • 0 Attachment
          --- In agile-usability@yahoogroups.com, Ron Jeffries
          <ronjeffries@X...> wrote:
          >
          > Interesting ... I might suggest references to Y.T.'s articles on ...
          > "running, tested, features", at
          > http://www.xprogramming.com/xpmag/AgileTopDown.htm and
          > http://www.xprogramming.com/xpmag/jatRtsMetric.htm


          My bad, thanks for the catch, it's already fixed. I had that
          reference in the Governance paper; didn't detect its absence here. I
          care about these things, so thanks again.

          >
          > Second, the article states that no self-respecting agilist will
          > permit iterations longer than two weeks. I think that at least some
          > of the Scrum people still have some remaining shreds of
          > self-respect, and the standard in Scrum is still a month.

          True, accepted, and still, I watch people's face as they take on
          board the idea that Scrum goes for a month ... so the song still
          holds.

          > Since he could lift me over his head, I'm not going to argue
          > self respect with Mike.

          I'm going to assume that Mike has self-restraint as well as self
          respect, so I'm happy to argue with him :-)
          (maybe not arm-wrestle for a beer)


          >
          > Third, the article suggests that short iterations are a matter of
          > pride. That is not the case in my opinion. AYMR, the original XP
          > project started with three-week iterations, tried one week, and went
          > back to three, having found it better. The /XP Installed/ book, I
          > believe, recommends three weeks. (ICBW).

          True about C3, but that was back in the mid-90s, recall. Note that
          the new XP calls for one week. You are now a fuddy-duddy like me
          (well, maybe not quite as fuddy duddy, because my default iteration
          length is still probably longer (for some definition of "iteration";
          and you are not really /like/ me; but the song under the lyrics still
          applies)


          > One of the interesting effects I've observed relates to this. I
          > encounter clients who, over the course of a month, have trouble
          > completing features. I tried recommending that they go to one week
          > iterations (I like the term "deliveries" very much, by the way, and
          > will start using it as soon as I step out of the thought here.)

          Please, only if they really deliver. If word-inflation moves
          into "deliveries", then I'm running out of words to use.

          >
          > An interesting thing happens when a team shifts to a one-week
          > iteration, with intention to deliver RTF: they have to focus. When
          > there are only five days till lift-off, it's easier to keep your eye
          > on the ball. I think that it's harder for Parkinson's Law to take
          > effect inside the shorter iteration. Whatever the cause, going to a
          > shorter iteration seems to help these teams learn to be DONE.

          ditto about actually delivering.
          Thanks
        • Ron Jeffries
          ... Well, that makes you a wuss, more than fuddy-duddy. Use of the term fuddy-duddy is self-incriminating, I suspect. ... Yes. I like delivery for that reason.
          Message 4 of 4 , Sep 10, 2005
          • 0 Attachment
            On Saturday, September 10, 2005, at 11:38:10 AM, aacockburn wrote:

            >> Third, the article suggests that short iterations are a matter of
            >> pride. That is not the case in my opinion. AYMR, the original XP
            >> project started with three-week iterations, tried one week, and went
            >> back to three, having found it better. The /XP Installed/ book, I
            >> believe, recommends three weeks. (ICBW).

            > True about C3, but that was back in the mid-90s, recall. Note that
            > the new XP calls for one week. You are now a fuddy-duddy like me
            > (well, maybe not quite as fuddy duddy, because my default iteration
            > length is still probably longer (for some definition of "iteration";
            > and you are not really /like/ me; but the song under the lyrics still
            > applies)

            Well, that makes you a wuss, more than fuddy-duddy. Use of the term
            fuddy-duddy is self-incriminating, I suspect.

            >> One of the interesting effects I've observed relates to this. I
            >> encounter clients who, over the course of a month, have trouble
            >> completing features. I tried recommending that they go to one week
            >> iterations (I like the term "deliveries" very much, by the way, and
            >> will start using it as soon as I step out of the thought here.)

            > Please, only if they really deliver. If word-inflation moves
            > into "deliveries", then I'm running out of words to use.

            Yes. I like delivery for that reason. When you say "deliver" to you
            wish to limit the term to those cases where the customer actually
            puts the software into production, or is it sufficient that they can
            run it and use it if they want to?

            Ron Jeffries
            www.XProgramming.com
            Hope is not a strategy. -- Michael Henos
          Your message has been successfully submitted and would be delivered to recipients shortly.