This wiki is locked. Future workgroup activity and specification development must take place at our new wiki. For more information, see this blog post about the new governance model and this post about changes to the website.
-- AndyBerner - 18 Jun 2009

Reviewing the initial estimate use case, the initial estimate provides a probability distribution of total duration and total effort, and there is some probability used for "nominal" values...an initial project plan will be based on that nominal, but the whole probability distribution is available (in some form) for the project or portfolio manager so they know both the nominal and how risky it might get.

Partway through, when the project is reestimated, we certainly may get a new nominal total duration. But the probability distribution has also changed. The work of the Tempo project from Rational CTO office/IBM Research is based on the goal that the risk, as measured by the probability distribution, should be getting lower over time if a project is running well. That is, if initially there was a 30% probability of the project taking 12 months, a (nominal) 60% probability it would take 18 months, but to get a 90% probability, you need to plan for as much as 36 months. If partway through I reestimate and now the nominal is 24 months, but there's a 30% probability is will take only 20 months, and a 90% probability it will take 28 months, I'm doing "better" because there is less risk in the schedule, even though it's longer than the original plan.

This affects OSLC interface: does the "reestimate" that is POSTED to the current estimate (or maybe POSTED to the list of estimates--we haven't set the interface yet) have to include just nominal values of the metrics, or new risk curves?

In discussions we have already had, your tools that do the reestimation currently do not ROUTINELY report a new risk curve, though that would be possible to do. So please comment on this:

Is the idea of keeping track over time of the risk, and managing to reduce that risk, not what most projects are looking to do? In your experience, why or why not?

Is there a difference in the probability distribution in the reestimate situation (we're this far along in the project, and we have this much left to do) vs. just doing an initial estimate of what's left to do as if that were a new project?

Galorath contribution

QSM contribution

Anyone else, please add additional document

Topic revision: r2 - 27 Jan 2010 - 13:56:48 - ArthurRyman
 
This site is powered by the TWiki collaboration platform Copyright � by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Contributions are governed by our Terms of Use
Ideas, requests, problems regarding this site? Send feedback