thanks for your response!
thanks for contributing the idea but I'd like to avoid this indirection - each indirection is one additional level of complexity.
yes this is true. A good example is the deployment descriptors of EJB (2 files, one container independent and one container dependent ;-))
See the deployment descriptor of the Hibernate cartridge, for example. It says there:
<stereotype name="Entity" /> <stereotype name="Service" />
This can be understood with one glance, you agree?
... but not complete. You don't put the tagged values, for example: finderMethod, primaryKey, etc. in this file, right? Adding a complete mapping would indirectly add every single tagged values, etc. in that file as well (not only the stereotypes), or am I missing something?:
<mapping name="entity" real-name="Entity"/> <mapping name="service" real-name="Service"/> <mapping name="finderMethod" real-name="FinderMethod"/> <mapping name="primaryKey" real-name="PrimaryKey"/>
One more argument: The cartridges, as they are today, are only examples of how cartridges can be. They should be adapted and/or replaced before a real world software project tries to do real world work with them. A skillful architect (member of the project team) should adapt them to meet the project's needs. Therefor, any additional complexity should be added by the architect, not by the AndroMDA team.
This is again true. But IMO, it would be also nice to have reusable cartridges. If I can reuse (not extending or customizing, just use them as they are) some cartridges from an experienced architect, that would be great. Maybe AndroMDA can build a marketplace for cartridges ;-) Okay, I know, all of you would say, where is the marketplace for EJBs? ;-) But I successfully reuse EJB components from LPortal in OpenUSS without changing anything, just have to rewrite the deployment desc., because I'm using JOnAS instead of JBoss. So, theoretically this could happen. And for this purpose we need to be able to reuse those cartridges without changing the name of the "stereotype" or "tagged values" in our model. Mapping is a standard approach.
Lower case is just the simple case. I can imagine that somewhere else you'll use <<type>> instead of <<entity>> or <<session>> instead of <<service>> but with the same semantic ;-)
Thanks and greets, -- --------------------------------------------------- Blasius Lofi Dewanto --------------------------------------------------- OpenUSS - Open University Support System http://openuss.sourceforge.net --------------------------------------------------- E-Mail : [EMAIL PROTECTED] ICQ : 39343280 ---------------------------------------------------
------------------------------------------------------- The SF.Net email is sponsored by EclipseCon 2004 Premiere Conference on Open Tools Development and Integration See the breadth of Eclipse activity. February 3-5 in Anaheim, CA. http://www.eclipsecon.org/osdn _______________________________________________ Andromda-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/andromda-devel
