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

3415RE: [scrumdevelopment] Re: SCRUM & Change / Defect Management

Expand Messages
  • Steve Bate
    May 4, 2004
    • 0 Attachment
      > > I'm curious why your developers are fine with daily updating of
      > > their remaining time estimates but they fight tracking actual time
      > > (again, for the purpose of calculating remaining time in the XPlanner
      > > context).
      > The issue is that the effort expended has no definitive relationship with
      > the remaining estimate to complete (ETC), the number can even be negative,
      > i.e. after a day of effort the task ETC may have increased.

      Hi Tom,

      Right, most teams use XPlanner to calculate ETC based on the task effort
      estimate and the actual time worked (ETC = estimate - actual). However,
      ETC in XPlanner is always equal or greater than zero. If the actual would
      exceed the estimated effort, a new (larger) estimate is requested. A task
      can be reestimated at any time. After reestimation, the ETC might be larger
      or smaller.

      > As we are
      > focused on the planning aspect of the process (will we complete
      > the selected work this Sprint?) the only thing we need to track is ETC,
      > not actuals. Psychologically it is easier to re-estimate a larger ETC at
      > the end of a days work than explicitly record the fact that your estimate
      > was wrong or that you haven't worked hard/smart enough.

      I wondered if that last issue might be one of the sources of resistance
      to recording actuals.

      Our team uses ETC to track intra-iteration (we're an XP team) status. We
      generally use previous actuals and estimation accuracy during our planning
      activities although a task that's been significantly reestimated during
      an iteration might trigger a related discussion in the standup meeting.

      For example, we learned that we tend to underestimate web development
      stories and the team determined it was because the functional tests were
      difficult to write. This motivated us to improve our web testing
      framework. Other stories would tend to run over because they depended
      on obtaining business information from our parent company and we
      were often passed to several intermediate people while obtaining the
      data. Once we noticed the trend of those stories exceeded estimates
      and determined why, we worked with the parent company to create more
      efficient ways to obtain the data we needed. I believe that having
      hard evidence of the impact of these inefficiencies decreases the
      response time in addressing them.

      Ken Schwaber discussed the pros and cons of tracking actuals in

      I agree with Ken that it's not a silver bullet. Still, we have found
      the extra feedback to be useful.

      > > Eventually I'd like to extend XPlanner so that the planning and
      > > tracking can be configured to handle a wide range of XP process
      > > variants. This (and a Scrum-specific web skin) should also make
      > > it an even better fit for Scrum teams.
      > Cool idea. Once I've finished my house move I would like to help (Though
      > that date is moving out two days per calendar day!)

      :) I completely understand. I'm constantly doing a time balancing act
      between my myriad activities and working on XPlanner. Although the Scrum
      skin could be created immediately, the advanced configurability will require
      some significant internal refactoring over several releases. OTOH, even with
      the current implementation, the skin could present a view that only tracks
      ETC. Like I said in a previous message, I've never used XPlanner in this way
      but there shouldn't be any major problems. If there are small problems,
      I'd be happy to make those changes to better support Scrum teams.


    • Show all 23 messages in this topic