On Fri, Apr 9, 2010 at 4:13 AM, Bob Tarling <[email protected]> wrote:
> This should be done in core ArgoUML though. What is the advantage in > having another project for this rather than contributing to the core > product? This is a university research project. If it produces workable results, the lessons learned can be applied to the core ArgoUML sources. I think it's great that it's using the ArgoUML sources - the more exposure and publicity the better. I think it's premature to consider any type of integration or use of the resulting code until more is known about the results. "Software Product Line" is a 10-year old buzz word out of CMU for concepts that have been important to software engineers for decades before it was invented - reuse and things that support reuse like modularity. http://www.sei.cmu.edu/productlines/history.cfm ArgoEclipse's approach to this problem is to decompose ArgoUML into more and more fine-grained plugins over time. These plugins can be mixed and matched with each other and with plugins from other providers (C++ tooling, Java plugins, etc). The common dependency that all these efforts have is cleaning the current design to untangle the current design and tease it apart into something that can be modularized. This is something that I would agree is best only done once, but it's a difficult and tedious task to do right and not impact other development efforts. I'm not sure I'd want a new developer attempting it on the production code base. Tom ------------------------------------------------------ http://argouml.tigris.org/ds/viewMessage.do?dsForumId=450&dsMessageId=2590622 To unsubscribe from this discussion, e-mail: [[email protected]]. To be allowed to post to the list contact the mailing list moderator, email: [[email protected]]
