Re: Alloy paper



Date view Thread view Subject view Author view

Frank V. Castellucci (frankc@colconsulting.com)
Mon, 31 Jul 2000 09:25:16 -0400


Alan Cameron Wills wrote: > > Daniel Jackson wrote: > > > substantial case for why it's worth considering a new language despite the > > standardization of UML. > > How should we go about getting the ideas in this contribution (and others like it) used by the programming community? > > I believe that: > * There is little hope of replacing UML --- there are too many people and businesses already signed up to it. > * Even if there were no UML, there would be so many contenders for "best" modelling language that we would spend all our time in debate about the virtues of them all, rather than getting down to using them. > > The strategy most likely to get some of our best ideas really used in widespread practice, is to gradually shove them into UML. My guess is that this will be true for about 10 years. > > I feel it is worth working on best-imaginable languages like Alloy, in order to experiment with and demonstrate techniques. But the next task is to migrate the ideas embodied in them incrementally into the (very imperfect) language that everyone is actually using. > > Alan Cameron Wills > Trireme International Ltd > Catalysis: objects, components and frameworks in UML Businesses used a plethora of other modeling languages prior to UML. If not for a few snafus on someones part, we may all have been using OPEN. UML was not brought to the mainstream attention because it was a coherent model, not by any means. A few identified names and a heavy marketing hand played a big part. I don't understand the "give over to the current winner" mentality you are prescribing, and must disagree with your avenue of approach. If the authors feel that what they have is superior, and accomplishes what we as a community REALLY require, then I say go get the ring. It will take time, granted, but that shouldn't stop anyone with a good idea. Now with the open source/free software tools arriving for UML, many of the project authors would really have no problem with adding a module to work with other methodologies and notation. For example, graphical representation and semantic validation/enforcement. Like XSLT, someone may want to focus on UML->Alloy or XMI->Alloy or RDF->Alloy or whatever can be done. But, it does mean work on someones part. And it really doesn't need to cost anything but time. My 0.02 -- Frank V. Castellucci


Date view Thread view Subject view Author view