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

Re: [XP] Re: My first trial run at the Planning Game

Expand Messages
  • Ron Jeffries
    ... This is good, slightly better than just killing the project. Looking at all the stories and selecting the best three months worth is pretty good, and at
    Message 1 of 242 , Dec 31, 2003
    • 0 Attachment
      On Wednesday, December 31, 2003, at 2:57:37 PM, jrb32002 wrote:

      > --- In extremeprogramming@yahoogroups.com, Ron Jeffries
      > <ronjeffries@X...> wrote:

      >> Right. There is at least one way to process six months of stories into
      >> three months of time that is better (i.e. never worse, often better)
      > than
      >> doing the most important ones first until done.

      > The suspense might well be killing some. Be interesting to see
      > whether my experience matches others'.

      > My first inclination is "kick it back to the Customer to figure out
      > which three months worth of stories is worth doing". If he can, do
      > them. If not, it wasn't worth three months of funding (much less six
      > months), so move on to something else which *is* worthwhile funding.

      This is good, slightly better than just killing the project. Looking at all
      the stories and selecting the best three months worth is pretty good, and
      at least has the advantage of making sure the customer understands the
      situation clearly.

      Your idea here uses an overall look at the stories, as does Joe
      Rainsberger's suggestion, which was to find a cohesive set of stories
      adding up to, say, 80% of the available resources. He was thinking that
      could be done without the information required by the Release Plan -- or at
      least that was the challenge he was interesting, but in fact the strategy
      seems to require all the stories and all the estimates, so I think it's
      equivalent for our purposes.

      Amir also suggested looking for key combinations of stories that might
      deliver more value when complete than when partially done. A good example
      that also requires an overview of the whole project in order to do it.

      The standing proposal from Mr Howell is to do the most important things
      first. Now one might object that the most important thing might take four
      months, but probably he meant the most important things in terms of highest
      return for lowest cost or something like that.

      Setting that aside, any of the ideas here, which involve doing an overall
      planning pass, will be at least as good, and usually better, than the
      proposed purely incremental approach. They accomplish this by focusing
      attention on constellations of features which have special value when done
      together (or deferred together).

      These are all on the right track. They all seem to be forgetting one key
      technique used in Release Planning, to get more features scheduled before
      the deadline. (And I'm not suggesting adding people, spending more money,
      or pretending we are faster than we are.)

      Hmmm?

      Ron Jeffries
      www.XProgramming.com
      The lyf so short, the craft so long to lerne. -- Geoffrey Chaucer
    • Ilja Preuss
      ... It probably also results in stories of different priorities. So you might happen to only implement part of the original story in those three months. Take
      Message 242 of 242 , Jan 19, 2004
      • 0 Attachment
        J. B. Rainsberger wrote:

        > Split the stories. The usual effect is that our fear-based estimates
        > are replaced with shorter, more realistic ones. This gets us more
        > features, because we reduce the effect of our tendency to
        > over-estimate.

        It probably also results in stories of different priorities. So you
        might happen to only implement part of the original story in those three
        months.

        Take care, Ilja
      Your message has been successfully submitted and would be delivered to recipients shortly.