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

3933[XP] Re: Frameworks

Expand Messages
  • Anthony Merton
    Apr 1, 2000
      > Introducing Frameworks or new technologies, Roguewave stuff,
      > MFC, COM or CORBA ... will require considerably study efford
      > of the teammembers. Understanding and getting familiar with
      > such a thing will need several days, or even weeks.
      > How can I can avoid to make the XP Process stuck on such matters?

      I hope I'm not OT, I just have the above paragraph as a start...

      First things first : only introduce those is you *have* to (that is, not
      because it is trendy, or that it "might" accelerate things after a few
      months etc.) Be focused on the project product, not on the means. Is such
      new technology an external project "constraint" ?. If yes, OK. If not,
      prepare yourself for a difficult fight.

      OK, so you should :

      - Have the team become familiar with the technology thingies during the
      initial exploration phase (where initial user stories are collected),
      developing prototypes with the new tools/framework.

      - Have the user stories estimated and risk rated taking this situation in

      - Have a load factor that initially takes into account the fact that the
      team is somewhat unfamiliar with the technology. You adjust it accordingly
      as time passes and the team become comfortable. Try to plan with a LF of 3
      or 4 initially and then bring it down to 2.5 gradually. You will have the
      Tracker tell you over time how to readjust it for the remaining work.

      - Accelerate and secure things by having on board a consultant who's "been
      there, done that", who will coach your people with the techie matters. Might
      be expensive, but you have to compare the cost with the benefits.

      At the extreme, you could (but have to explain it that to stakeholders)
      *plan* to throw out everything you have developed after a few weeks / months
      and start over. I say at the extreme, but this is really what one should do.
      For instance, the second release adds no user values, and is a total
      rewrite. Now this would be opposed to the XP planning game! So there is the
      approach to refactor totally your system over quite a few development
      iterations. This is more complicated, but has to be done.

      If you can explain that to customer / management and they agree, you'll be
      well off. Ask one of the senior project managers in your company how to
      manage the customer and the management in such a situation. Here is the
      point mentioned above : if introduction of those technologies is an external
      project constraint, then you have a zillions good reasons to bring forward
      to customer and management that it will take more effort they would like it
      to. It's all about if they're serious about risk or beleave in fairy tales.
      If it's not an external constraint... ouch, it might be difficult to justify
      the extra effort, but do plan it anyway, and be open about it, or just drop
      it !

      Finally, take only the *best people* you have on board.

      Grand Finale : whatever you do : PLAN for it, or it will bite you !

      Does this help ? Wow, this was a long one. Did anyone really read until here
      ? :-)

      Anthony (project manager)

      > Christian Gantar
    • Show all 3 messages in this topic