> I shall eventually copy UML shapeset, create another called ERD, and > change the syntax to be SQL specific so there doesn't have to be a > vocabulary mapping anymore.
It just occurred to me that since UML and ERD will be 99% similar with mainly only a vocubulary difference, that it might be wise not to fork UML & ERD. What would you list subscribers guess would be the best method for *NOT* forking the UML shapeset, and renaming the labels & values for various checkboxes, text boxes, etc? Here's a quick example: On the Class shape, there is a checkbox on the Class tab called "Abstract" which for ERD I would call "View." Also, under the Attributes tab, there's a drop-down labelled "Visibility" which I would relabel "Column Type" and the values would be not Public/Private/Protected/Implementation but Normal/PK. Being a non-coder, my guess is there needs to be a pre-processor that generates both class.c and dbclass.c which are different in those subtle ways...? Or is it quicker/easier just to fork the shapeset and try to keep the updates consistent across both shapesets? -- Tim Ellis Senior Database Architect Gamet, Inc. _______________________________________________ Dia-list mailing list [EMAIL PROTECTED] http://mail.gnome.org/mailman/listinfo/dia-list
