Modeling PMC Meeting, 2006-06-20
If you've got any specific items to add to the list of topics for tomorrow, please reply.
I have (again) tried coming up with a structure for the Modeling project, and propose this be our primary topic for tomorrow. We'll need to resolve this soon in order to move forward with provisioning, etc. I've attached a slide (with notes) that Ed and I feel is a good starting point for the discussion, so please review at your convenience.
- Richard Gronback
- Ed Merks
- Kenn Hussey
- Fred Plante
- Sebastien Demathieu
- Jean Bezivin
The primary discussion point of the call was regarding project structure within Modeling. A number of suggestions were made in reference to the initial proposal, sent earlier (see below). An updated slide with the changes discussed on today's call is attached.
It is planned to have this issue resolved no later than our next PMC call (July 18th) so that provisioning can commence ASAP.
- Net4j/Teneo components for EMF are to remain within EMF, although not indicated in slide (fixed).
- The Eclipse Modeling Tools (EMT) project proposed to house modeling tools was generally accepted, while Kenn thought Model[ing?] Development Tools (MDT) was a more appropriate name (most agree, it seems as it fits nicely with JDT/CDT). Fred expressed some concern with including 'Tools' in the
name. It was suggested that Kenn would take leadership of such a project, as it would initially contain UML2 and related tooling. The inclusion of "industry standard" in the name was proposed, and seems fine except for the case in the where a modeling 'tool' in the future might be added but not follow an industry standard.
- Kenn also pointed out that MDT (or whatever) should be the home for the OCL project, as it is a spec found within UML2 according to the OMG's organization. A discussion of the project's namespace followed, as it's currently org.eclipse.emf.ocl, while it would be more appropriately found in org.eclipse.ocl (according to convention). Unfortunately, we've just agreed in the Release Review that OCL was to exit incubation in its current state, which leaves the next major release version as the time to make such a refactoring.
- The MXF (Model Transformation Frameworks) project was thought better split in two by Fred. This was generally accepted and the project names are proposed to be M2M (Model to Model) and M2T (Model to Text). (I guess MXM and MXT could be alternatives that save the 'X' ;-)
- Fred proposed, as was suggested in the slide, that the GMF and TMF projects be combined into one (MXF represents an analogous combination) with "notation" in the name. It was agreed to keep them separate, which is in keeping with the proposed separation of MXF into M2M and M2T.
- As JET seems to have the most progress toward the formation of a standalone project for M2T, it is believed it would form the initial M2T component, perhaps with Paul Elder as lead. There is also the proposed MOF2Text project (still not public) and the xPand component within GMT to consider.
- We need to get feedback from Jean regarding potential rename of GMT to indicate it's nature as both technology incubator and research project. Currently, it's Generative Modeling Tools and one suggestion is Generative Modeling Technologies. This seems to imply it is only for "generative" tooling, however. It will be left to Jean to discuss impact of this or alternative namings, as well as feedback regarding the whole nature of this change (additive) as he is the project lead.
- A final note is in reference to switching to Subversion as the repository provider for Modeling. It was agreed that it is best to wait for resolution on a proposed new build facility for Eclipse, as there will be a necessary integration required. Also, the existing scripting used and hopefully to be
leveraged throughout Modeling works with the existing infrastructure.