RE: Sets and bags || UML and PUML going astray



Date view Thread view Subject view Author view

robert france (france@CS.ColoState.EDU)
Thu, 25 Jan 2001 19:25:41 -0700 (MST)


>if you give a modelling language a low-level, implementation-based >semantics, it's no longer appropriate for abstract modelling. Agreed; but is this really true of the UML or are you referring to some of the *desired* interpretations given on this list? The more I see postings on semantics, the more I see the potential for UML profiles. For example, a UML profile for Analysis can assign appropriate semantics to the UML constructs. I really would like to see (p)UML'ers examine the role of profiles in defining semantic variations of the UML (thus making the UML a family of languages). This discussion on "semantics" of association reminds me of similar discussions on aggregation (no, let's not go there again!). An important lesson learned from that experience (at least on my part!) is that the folks who matter (the developers in the "trenches") will use a "semantics" that they are comfortable with and fits the problem context in which the UML is used. Personally, I prefer the set-based semantics for associations (an association is a set of tuples ...), and the notion of links having identity is somewhat lost on me, but all that means is that I've not come across a situation that required me to consider this; until then I'll flow with the set of tuples interpretation ... Ofcourse I'm saying all this without reading all the preceding emails carefully; a dangerous thing to do ... OK, I'll go back to the sidelines now! Robert ==================================================================== Robert B. France, Assoc Professor | Tel: 970-491-6356 Computer Science Department | Fax: 970-491-2466 Colorado State University | Email: france@cs.colostate.edu Fort Collins, CO 80523 | www.cs.colostate.edu/~france/


Date view Thread view Subject view Author view