Re: Hello
Jim Sabatke <[email protected]> Fri, 17 Nov 2006 05:08:55 -0600
| Newsgroups | gmane.comp.kde.kplato |
|---|---|
| Message-ID | <[email protected]> |
Dag Andersen wrote: > Fredag 17 november 2006 10:00 skrev Jim Sabatke: >> Dag Andersen wrote: >>> Mandag 13 november 2006 19:49 skrev Jim Sabatke: >>> (...) >>> >>>> Even GANTT charts have limited use IMHO. The best charting available is >>>> to display Budgeted Cost of Work Scheduled vs. Budgeted Cost of Work >>>> Performed (Schedule Variance); and Budgeted Cost of Work Performed vs. >>>> Actual Cost of Work Performed (Cost Variance). >>>> >>>> Those two graphs will give the project team and executives a very good >>>> view of how the project is actually proceeding. >>> How many datapoints do you think we typically need to plot? >>> I'm asking because it would be nice to use kchart for this but there is a >>> practical limit to how much data it can handle. >>> >>>> Estimated Cost and Time of Completion are also useful. >>>> >>>> Jim >>> (...) >> You actually don't need that many data points. The graphs show >> incremental progress, so depending on the size of the project, daily or >> weekly data points are sufficient. The idea is that there should be >> enough data points to accurately convey the real status of the project. > Diplomatic :) > I'm playing a little bit with kchart and I'd say *much* more than 20 values > (per curve) and we get into problems. First with text along the x-axis (one > text per value), then with performance. It's made to resize to it's window, > (no scrolling) so the largest chart would be a full screen one. We can always > put it in a scrollview though, if needed, put then performance comes into > play. Hmm, it's not an exact sience... > > OTOH whats the granularity of the performance data? We used to enter data > weekly for projects of 3 months to approx a year, so it would be hard to show > accurate data for smaller intervals. >> Jim >> _______________________________________________ >> kplato mailing list >> [email protected] >> https://mail.kde.org/mailman/listinfo/kplato > I think that weekly data points are pretty much ideal. That's the granularity I usually used when I was using CA Superproject (now defunct), which, btw, would have made a good model to copy for kplato. Jim