Dear Wayne,
Before starting that proposal we had of course a discussion whether this should be separate project, and we agreed on putting this into a new project due to the following reason: The fact that UML2 provides facilities to implement profiles does not necessarily mean that profile implementations have to b hosted under that umbrella, even though UML2 provides the fundamental technical and semantic basis. By that logic almost any EMF-based project could be based in the EMF project itself. The UML2 project copes with implementing and maintaining the UML2 specification, a comprehensive and demanding task that should not be complicated by dealing with sometimes comprehensive and demanding profiles (like SysML or MARTE which have nearly the same number of pages than UML itself). This would probably far too much for the UML2 project. In addition, even if those profiles only evolve once a year there can be actually a lot of profiles being hosted, so that it would make definitely sense to organize them in a separate project rather than in hosting them somewhere else in MDT. It would just represent the next (some people already arguing a “necessary”) step for the MDT project in order to prevent redundant and maybe incompatible implementations of the very same profiles within the Eclipse community.
> Having said all that, the "Initial Committers" section is really an "Initial Contribution" section. Who are the initial committers? We also need a >"Why Eclipse?" section as defined by the proposal template.
I’m going to split and clarify these sections soon, i.e. within the next two days.
Best regards,
Marc-Florian
For completeness, a project is "just a container project" by convention only. The formal notion of "container" was removed from the EDP some time ago. It is perfectly natural/reasonable for a mid-level project to have code/resources and committers. Whether or not that makes sense here is a different question.
If creating a whole separate project is the right thing to do, then we should do that. My primary concern is that we have a lot of modeling projects with 1 or 2 active committers and I often hear complains that maintaining a project is "too hard" or "too much work". By creating a new project, we're adding to that work.
I'm also not entirely comfortable with a project made entirely of components that "usually evolve once a year". Though, I guess that Orbit sets a precedent here.
Having said all that, the "Initial Committers" section is really an "Initial Contribution" section. Who are the initial committers? We also need a "Why Eclipse?" section as defined by the proposal template.
Wayne
On 05/29/2012 04:12 PM, Kenn Hussey wrote:
Well, this functionality is (mostly) independent from the UML metamodel per se and would have a different lifecycle and set of committers... and the MDT project is just a container project, so I'm not sure it would make sense for it to be hosted at that level.
Kenn
On Mon, May 28, 2012 at 4:11 PM, Wayne Beaton <emo@xxxxxxxxxxx> wrote:
Hi Kenn.
Does this really need to be a separate project? We can set up all the things you list, website, mailing list, etc. without creating a whole new project.
Does it make sense to consider extending the scope of UML2? Or might it be a resource that maintained by the MDT project?
Wayne
On 05/22/2012 04:14 PM, Kenn Hussey wrote:
The Modeling project would like to propose creation of a new "UML2 Profiles" subproject of MDT. The proposed project is about providing implementations of "standard" UML profiles and a repository in/from which to discover/install them.
--
Wayne Beaton
The Eclipse Foundation
Twitter: @waynebeaton
Explore Eclipse Projects
_______________________________________________
modeling-pmc mailing list
modeling-pmc@xxxxxxxxxxx
https://dev.eclipse.org/mailman/listinfo/modeling-pmc
--
Wayne Beaton
The Eclipse Foundation
Twitter: @waynebeaton
Explore Eclipse Projects
