[cap-talk] Re: Question on web calculus
lefevrol at yahoo.com
Fri Feb 17 17:32:12 EST 2006
> Perhaps you could also clarify what you are looking for in a
> "_theoretical_ analysis".
That's a fair question but I will note that you are the one who made
claims to theoretical analysis and that you no more defined what you
mean by "theoretical" analysis than I did.
What I have in mind when I hear "theoretical" in a CS context is the
kind of proofs you find in papers published in Theoretical Computer
Science (http://www.sciencedirect.com/science/journal/03043975) or even,
not to put the bar so high, in Hoare's CSP book. Crudely put, theory ==
mathematics of some sort. I think that is a fair definition: the journal
in question goes 30 years back and is one of the pillars of the field;
the fact that I don't understand 99% of papers in there (pre-empting
another swipe from you) does not detract from that: you cannot redefine
the meaning of theory in the CS field.
Now, what you describe in your second paragraph is reasoning but it
doesn't strike me as particularly theoretical. Notwithstanding your
swipe at non capability-aware programmers, all programmers have to
reason about their programs and some of that reasoning has a similar
flavour to what you described. For instance when reasoning about memory
leaks they are also reasoning about reachability attributes and the
whole object reference graph yet I have not heard anyone claiming that
in so doing they were doing theory.
> Since you did not attend the presentation, your assertion is unsupported.
But most people did not attend your presentation either! As a proof that
Waterken lends itself to "theoretical" analysis, it is only valid to
the few who were in the audience on that day.
Now, before you fly off the handle, let me conclude by saying that I
don't particularly care whether there is much "theory" (in the
lots-of-equations sense) behind capability-based systems like Waterken.
Indeed what appeals to me about this corner of software engineering and
in this forum is that people here clearly try to think rigorously but do
not act like you are disqualified if your argument does not include
several pages of equations: instead, they strive to reason in plain
english. It is with this in mind that your recourse to the shibboleth of
theory strikes me as so unfortunate.
And with that I will shut up.
More information about the cap-talk