[news.eclipse.tools.hyades] Re: Model group summary meeting April 16

["Harm Sluiman" <nobody@xxxxxxxxxx>]

All the sub-model group leads joined this call.

We discussed the need for a test execution history model, and concluded that
a sub-group would be appropriate. I will post a call for participation, but expect we
already have some people keen to cross pollinate. Marcelo has already offered to
take the group lead role.
This group will have to look into how it should link to the test definition model.
In some previous work the notion of a test instance was used to link a definition
with a configuration and any execution histories that resulted in runs of that set.

The Statistical model has one issue left to resolve but that will not make the drop
for next weeks build.

Test and Trace models both have some incremental changes coming in the next week.
The trace model has enough info agreed that there is already a UI and collector for
the basic trace concepts. The trace model group is closing on the "annotations" of 
the various data collected and will them move to the heap model work.

The Test model is about at the point where the first definition editor samples can
be composed and that work will be under way shortly to drive back iterative requirements.

We also discussed that no matter how much commentary we put in the models there
will be semantic requirements placed on the data providers and model users that need
to be identified very clearly. For example the Object id in the trace model must be unique,
and documenting that in the model will not jump out at a profiling agent implementer. It
was decided that this should be part of the conformance/compliance specification for the 
model(s). This will have to become a working document initially and we will evolve it
live on the Hyades web site.

Each group is going to continue evolving the documentation they provide for their models. For
Hyades 1.0.0 which is targeted for June 2003, the structural parts of the models will have to be
settled down by early May at the latest, and a significant amount of documentation will need to be 
created to match.