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



Date view Thread view Subject view Author view

stuart kent (stuart@mclellankent.com)
Thu, 25 Jan 2001 09:42:41 +0000


Daniel > i disagree. one of the most valuable ideas in OMT, now apparently lost in > UML, was that of an "analysis model" that abstracted away from > implementation details. if your modelling language is built around > programming language notions, you'll never be able to use it in the stages > where it's most valuable -- long before you're worrying about code. I don't think this is lost in UML. I agree that the way most people tend to use UML is as a language to visualise designs that are not far off the code. If one restricts oneself to class diagrams (leaving out some implementation-oriented stuff), OCL to write invariant constraints and pre/post conditions, you get something that has much in common with your language. Packages are also important for breaking up large models, though these are confused in UML. I have used such a subset e.g. in modelling telecomms networks and building analysis models. This notation set can be extended with state diagrams (interpreted as an abstraction of state characterized by class diagrams and an abstraction of pre/post behaviour), which can be a useful way of looking at a system. What has been missing is a precise definition of this subset and tools to support it - though the latter is beginning to be addressed by tools such as the USE tool from Bremen. I think there is a chance with the UML 2.0 effort that some improvements can be made in this area. In particular, the idea that UML is actually a family of languages, means that we could have a family member or members suitable for modelling at a higher level (analysis, requirements, business), as well as members that are more suitable for implementation modelling. In the end, I guess, there are two approaches. Invent new languages or try and improve existing ones, especially ones that already have widespread acceptance. My leaning is towards the latter approach. Indeed, the work that some of us in pUML are trying to do is to move towards an approach to language definition which allows languages to be adapted and evolved in families (and, of course, these definitions must be understandable by tool implementors and the like). If UML gets defined using such an approach, then I think that it will be much easier to propose incremental improvements and changes. Stuart -- ukc home - http://www.cs.ukc.ac.uk/people/staff/sjhk sse research group @ ukc - http://www.cs.ukc.ac.uk/research/sse puml group - http://www.puml.org/


Date view Thread view Subject view Author view