[cap-talk] Question on web calculus

[Olivier Lefevre]

I read the threads on this mailing list (well, some of them) and the 
slides from Tyler Close and Alan Karp. Where it really clicked for me is 
when Alan uttered the word disaggregation. If I understand this 
correctly, the key insight would be that by adding extra nodes (i.e., 
URLs) so that there is now a 1-to-1 correspondence between {user, 
service} pairs and nodes, the problem of offering access guarantees is 
turned from one of verifying _behaviour_ (i.e., what happens when 
multiple users and services are multiplexed at one big portal) to one of 
analyzing the _topology_ of a graph. The latter is quite tractable: we 
can tap a very large and mature body of graph algorithms, whereas the 
former is quite hard: it takes you into formal software proofs, a much 
more esoteric topic and one that is perhaps not ready for prime time.

Did I get this right or am I off the mark?

-- O.L.