[cap-talk] Re: Question on web calculus

[Olivier Lefevre]

> Yes, you got that part right.

Thanks. Now, I don't want to reopen that nasty thread on LC, so don't 
read the footnote [*] below if you are still smarting from it, but it 
seems to me that the insight above is quite independent from any 
software. Perhaps it should be allowed to stand on its own: both the 
software framework and this reachability insight would gain (IMO) from 
leading separate lives. Just a suggestion.

Regards,

-- O.L.

[*] For the record, though, I am with Ka-Ping Yee on this: the LC is too 
general, too foundational for its invocation to be really useful: it may 
trigger a ha-ha moment in those with that background but I would contend 
that it does not help us _understand_ the other thing. For instance if, 
as some have argued in the other thread, even Java is drenched in LC 
goodness, then it becomes impossible, from looking at it from that 
angle, to understand how it differs from, say, Lisp (whose ties to LC 
are arguably much closer) and why you'd use one rather than the other.

As for the claim that the LC connection somehow vouches for the general 
soundness of the approach (which seems to have been the take-home 
message in Alan's slides), if I were one of the CIOs in the audience, 
that would not impress me, on the contrary such hand-waving would alarm 
me: there might very well be a connection but for all they know (and 
they don't know _you_) you might have gotten it all wrong, too, and 
misused those LC concepts. It's a bit like people claiming their product 
is secure because somewhere, somehow, they use RSA (or whatever). There 
is (unfortunately for busy CIOs) no substitute for actually reading the 
papers and examining the claims.

On the other hand, pretty much everything and anything seems to be 
labeled a "calculus" those days, so why not a "web calculus" indeed? 
_That_ does not bother me: at least it's a familiar term.