I realize now that as I (re)wrote the problem statement I ended up
with a general problem statement for Web transport architecture; this
is clearly too large to handle as a single problem, or by this group.
I believe the area we can make progress on most quickly is the
client/server protocol and transport between client and server;
i.e. very close to what Simon's proposal is in scope. We need
to keep our eyes on the prize of much better performance for the Web
as it currently exists, and better behavior over low bandwidth/high
latency local loops.
We do have to keep in mind that what we are
doing is just one part of the overall architecture, however,
and that large problem is what the problem statement ended up describing.
We have to think about these issues enough to make sure we don't miss anything.
Doing the client/server protocol without any thought to
the larger problem would likely result in missing things in the design
of this protocol.
I'm going to to be making a presentation to the IRTF's End to End task
force on the general problem early next month to try to stir up action
in the research community on the larger/largest issues (server/server
caching and notification), to take the Web to full scale and load.
I'd like to suggest people look at Web Forum and see if you like it;
so far, only Jeff Mogul and Simon Spero have signed in. If you
don't like it, we can just use mailing list technology.
See http://webforum.w3.org/wf/w3c/dispatch.cgi, and either attempt to login
or take the Registration Desk hyperlink to sign in.
- Jim