[news.eclipse.tools.hyades] Re: Model team meeting March 5

["Harm Sluiman" <nobody@xxxxxxxxxx>]

model team March 5 meeting minutes

Attending call:

Mike, Richard, Barbara, Marius, Joe, Serge, Julien, Gian Franco

(Mike, IBM, Rational, SAP)

Each of the model subgroups reported how far they went in the

past week. Progress has been made on all fronts, and it was

decided that we would be able to publish a first cut of

every sub model in about a week. The material will focus

on a key point in each area. We settled on a notation

convention in the model diagrams and that will be explained

in the Main page of the diagram.

it was agreed that the common base event model being worked on

by various GRID projects will be used as the base for our

log model.



Test as reported by Joe

The test group has focused on the linkage and re-use of the

UML2 Test Profile model, and how it would manage any dependencies

on base UML2 concepts.

The conclusion is to take a copy of the U2TP model as of the

March 3 OMG draft, as well as create a model that isolates and

reflects the dependencies on UML2. This will let the project

proceed with the test model, and also minimize the changes

needed if linkage to a real UML model is done down the road.

The copy of U2TP will be published on the site, but only for

reference purposes. Our actual model is expected to be significantly

larger that the U2TP model and will extend it with additions of members

and association versus and removal or semantic change to the model.



Trace as reported by Allan

The focus is to cover "continuous stack trace" of Java initially and

then add heap analysis and distributed support, followed by

multilingual support.

The team expect to be ready with this in time for the March 12 call.



Behavior as reported by Harm

After the various other sub-teams met and concluded their need for

a common behavior model was limited, it was decided this team would

go on hold until the test and trace model team define their exact needs.

At that time this team will reconvene and reconcile the models in this

common area.



Statistical as reported by Richard

The statistical model team has quietly met and basically settle on

50% of the model. It will be published along with rest.