Hi Wouter,
On Fri, 2005-02-04 at 14:34 +0100, Wouter Zoons wrote: > > On a related note, is there anyone interested in the AndroMDA community > > with working with ArgoUML to get them moved over to MDR? They seem to > > be a little short handed, and I think it would great benefit to AndroMDA > > and the open source community at large if a developer could do Model > > Driven Development with purely open source tools. An additional benefit > > is that I think could be some very nice synergy between ArgoUML and > > AndroMDA. > > > > [WZ> ] I would help out, but the problem is that ArgoUML only supports a > subset of the UML (1.3) specification .. that subset is so small that I > personally have zero benefit from it if it would be compatible with > AndroMDA. > > Most of the stuff I'm missing is in the activity graphs: parameters on > transitions, deferrable events, tagged values on *any* model element > (ArgoUML only supports a single tagged value, and I don't think you can add > such as tagged value on any random element), You are right about parameters on transitions and deferrable events. However, ArgoUML does support multiple tagged values almost everywhere. What it doesn't support is multiple stereotypes, which is consistent with the UML 1.3 metamodel. My point is that if an concerted effort was made to create a MDR based backend for ArgoUML with a UML 1.4 or UML 1.5 metamodel, the current ArgoUML developers could easily modify the existing property editors to support the things that AndorMDA users and developers want in a case tool. Then Andromda would not be held hostage by any commercial case tool. >From what I can see, much of the ArgoUML development over the past year has been to decouple the interface from their current UML respository, NSUML, which is a dead project. FWIW, I think the ArgoUML community would be happy to add features that AndroMDA needs if we could help them get to MDR and UML 1.4 or 1.5. > > > I would be willing to devote some time, since I am somewhat comfortable > > with MDR, but I don't have the time to do it alone. > > > > I know there was some work on creating a UML 1.3 metafacade > > implementation with ArgoUML in mind, but I think that approach is > > inherently flawed because of the limitations of UML 1.3 with respect to > > stereotypes and tagged values. > > > > [WZ> ] I agree > > > I think it is very unfortunate that an open source project forces you to > keep upgrading a commercial tool in order to use it. > > [WZ> ] it's indeed unfortunate, but it's not exactly the AndroMDA project > that forces us into this position, it's the sloppy and incomplete > implementation of the UML specs in those CASE tools .. (see: > http://team.andromda.org/docs/case-tools.html) > > Let me just note that if you are not using bpm4struts you can perfectly use > Poseidon to model for the spring/hibernate/ejb .. cartridges. So somehow the > way cartridges are implemented makes it more difficult to find tools with > the proper support. This is true, but you cant read in the cartridge models into Poseidon because of the way they are stored by Magic Draw. I can read the UMLMetafacadeModel into Poseidon and work from that, but I cant load the SpringMetaFacadeModel on top and look at that. Poseidon doesn't merge share models produced by MagicDraw, for some reason. If the cartridge models were made available in a non-shared fashion, they could be read into Poseidon. With the current situation, a person has to purchase Magic Draw 9.0 in order to open up the cartridge models. > > Mind you, I guess we could make more tools compatible, but the price would > be that users would need to model their applications in a counter-intuitive > way, with a lot of redundancy, not to mention the maintenance that would be > a nightmare. I completely agree and I am not proposing that. > Anyway, I guess all CASE-tool vendors offer different solutions based on > what they assume the public wants, .. I personally want to see support for > everything, and there's currently only one tool I've found that offers > everything I need. My proposal is to create a choice, with one of them being open source. I think there could be a lot of synergy between the two communities if the ArgoUML and AndroMDA tools could work together, and getting them on MDR with UML 1.4 seems to be the most natural way to do it. cheers, roy ------------------------------------------------------- This SF.Net email is sponsored by: IntelliVIEW -- Interactive Reporting Tool for open source databases. Create drag-&-drop reports. Save time by over 75%! Publish reports on the web. Export to DOC, XLS, RTF, etc. Download a FREE copy at http://www.intelliview.com/go/osdn_nl _______________________________________________ Andromda-user mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/andromda-user
