Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!usc!snorkelwacker.mit.edu!bloom-beacon!eru!kth.se!sunic!mcsun!ukc!ox-prg!culhua!mike From: mike@prg.ox.ac.uk (Mike Spivey) Newsgroups: comp.specification Subject: Re: VDM vs Z notation? Message-ID: Date: 17 Jun 91 08:54:20 GMT References: <48050001@hpindda.cup.hp.com> <48050004@hpindda.cup.hp.com> Sender: news@prg.ox.ac.uk Organization: Oxford University Computing Laboratory, UK Lines: 30 In-reply-to: collin@hpindda.cup.hp.com's message of 13 Jun 91 23:03:34 GMT In article <48050004@hpindda.cup.hp.com> collin@hpindda.cup.hp.com (Collin Park) reports that Sean Matthews (sean@castle.ed.ac.uk) writes: Z ... has a fairly good semantics produced by Mike Spivey, once you get over the notion of a semantics for Z in something that looks so much like Z. But Z could be accused of having other problems: for instance some aspects of the notation (the part invoving decoration of schemas and variables, is to my mind, a bit dodgy --- to say the least). I don't understand how Z can have a "fairly good" semantics, and yet have such basic parts of the notation be "a bit dodgy --- to say the least". In fact, it is a major goal of the semantics to clear up exactly what is the meaning of decoration and other operations on schemas. Perhaps Sean Matthews would explain what he thinks is the problem here? Later, he says, Also, VDM has better facilities for structuring very large specifications --- the notion of variable scope is a bit woolly in Z. It is there, but it is not so clear, and the user has to do more explicitly. In fact, ZRM as described in Jones' books has NO facilities for structuting large specifications (though some were considered by the standards effort). Again, I'd be grateful for some indication of how a "fairly good" semantics can leave the notion of variable scope "a bit woolly". -- Mike Spivey