[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
The Laws of "object" programming
One of the current limitations of distributed object programming models is the
lack of "reasoning" models available. In particular, there are loads of Java
based programming tools out there now (Concordia, Aglets, Voyager, Odyssey
agents, JavaBeans) which facilitate distributed programming - but none of
these give you any "framework" within which to reason about your programs,
their behaviour and their correctness. They are essentially engineering
I have been working over the past three weeks with Mani Chandy in Caltech
talking about these issues and I feel the the "java-threads" and CSP channels
model gives the object community a sound model upon which they can add some
reasoning rules. There are two distrinct sets of rules I am interested in:
firstly, the rules which facilitate reasoning and equivalence statements; and
secondly, classic structures, or patterns to use the OO communities phrase,
used to solve problems.
I am going to draft over the next few weeks some ideas about translating some
of "the rules of occam programing" into "the rules of object programming". Now
some of these may be intracable because of semantic incongruencies, but I have
an intuituve feeling that many of them will readily translate.
I would be interested in everyone's feelings about this approach,salient
technical insights, and even suggestions. If anyone wants to talk a little
more about this, just drop me a mail!