Re: Alloy paper



Date view Thread view Subject view Author view

Joaquin Miller (miller@joaquin.net)
Fri, 28 Jul 2000 21:07:54 -0700


Robert is, of course, right. But i can tell you that there are some people working on the UML who have made comments like "Where are the formalists when we need them?" The biggest practical problem with informal systems is that they allow us to argue endlessly. The one practical problem with formal systems is, of course, that they have no meaning at all. But the terms of a modeling language in use do have meaning. It is easy for a human observer to learn the meanings by watching. I feel the two camps Robert describes are both off target. What is needed is work along the lines of what was done with RM-ODP. The UML needs a clear and clean specification in ordinary, informal language, which specification is solidly based on a formal foundation. The thing to do is to develop a formal system, beat it into shape, then translate it into a set of natural language concept "definitions." These so-called definitions will be like those in a dictionary. Either there will be circularities or the definitions will use undefined terms, or both. [That's as it should be. Remember, as Bunge said, if there is one thing that twentieth century philosophy has taught us, it is that we can not define everything while avoiding circularity.] Some of the UML 2 workers will welcome help from formalists. [We also need to provide as much compatibility with UML 1.4 as reasonably can be provided. It's not nice to abandon your installed base. This can be done, but not if it is not attempted.] Cordially, Joaquin Miller Chief Architect Financial Systems Architects mailto:joaquin@acm.org San Francisco phone: +1 (510) 336-2545 fax: +1 (510) 336-2546 PGP Fingerprint: CA23 6BCA ACAB 6006 E3C3 0E79 2122 94B4 E5FD 42C3


Date view Thread view Subject view Author view