Topic:
Simplification of Hyades statistical
tables.
Attendees:
Curtis d'Entremont
Eugene Chan
Wayne Ho
Andrew Kaye-Cheveldayoff.
Time:
Wednesday, July 28, 12:00pm (noon) EST.
Notes:
There was agreement on the design for
the simplification of the statistical tables. New icons to be requested
for views (memory statistics, execution statistics) and toolbar (different
levels). Icons and labels to be same/similar to coverage statistics view.
There was some discussion on the sequence
diagrams as well, and whether we should keep going with the splitting from
one sequence diagram into each type of interaction (host interactions,
class interactions, ...). This idea was put on hold for now. We are instead
considering a similar approach as with the statistics views with the different
levels of information. We would have two interaction views, one for log
interactions and the other for all the trace interactions (would need to
come up with a good name for this view). An important question raised is
whether we need to have two diagrams open at the same time for comparison.
Another important consideration is a "drill-down" capability
for the user. For example, to be looking at host interactions and want
to drill down to see the thread interactions for that host, then to the
class interactions, etc. Switching to a lower level would show too much
information; from all the children, instead of just the one we want to
drill down to.
There were no conclusions on this issue;
it will be discussed further in next week's UI meeting, to be scheduled
soon. We would like to include some of the sequence diagram folks from
Toulouse in this meeting and try to find the best approach.
Thanks,
Curtis d'Entremont
Problem Determination Tools
IBM Toronto Lab