At 15:41 +0100 20/10/97, Greg Michaelson wrote:
(rearranged by me for dramatic effect :-)
>But you should look at the Jones and Hayes paper from Software
>Engineering Journal called something like "Executions are not (always)
>executable" to discover why persons of good taste in the formal community
>will no longer speak to you if you take this any further...
You should really put a smiley after that
>Surely the very suggestion that formal specs can be translated into anything
>is heresy, unless you're working with an executable subset.
Au contraire - this seems to be the orthodoxy. Lost of people in formal
methods now seem to be talking about executable subsets, and the need to
"animate" specifications - for exampe a tools vendor that is very well
known in the formals methods area was demonstrating ideas such as dynamic
linking of specs
to graphical animation tools, as well as the generation of C++ (Wot ? not
even a nice functional language 8-O )
>But you should look at the Jones and Hayes paper from Software
>Engineering Journal called something like "Executions are not (always)
>executable"
Please - can we have the full reference - Year, Volume no ,etc ... ?
Lots of smileys - its too early in the morning to get serious
lotsOfSmileys = ":-)" : lotsOfSmileys
_____________________________________________________________
Andrew Butterfield, Location: LG.19
Dept. of Computer Science, Tel: +353-1-608-2517
O'Reilly Institute, Fax: +353-1-677-2204
Trinity College,
Dublin 2, IRELAND. mailto:[EMAIL PROTECTED]
URL: http://www.cs.tcd.ie/Andrew.Butterfield/
