2 MS people will attend Hawaii meeting (me and Josh cohen)
We will attempt to have a preminary document available before the next
teleconf. We will have a fuller document available before Hawaii.
The consensus within MS was that, since there is already a separation
between model/semantics and protocol, the remapping should be relatively
painless.
Specific technical points:-
XML parser: There should be no problem in making a parser in < 50k. We have
both java and c code that we might be able to share with people. A subset is
possible - it looks like the final XML spec will allow for that.
Application/ipp: The immediate feeling was that instead of application/ipp
we would use multi-part. This was after 10 seconds of detailed analysis of
the issue :-). The PDL would be carried as a separate opaque BLOB as it is
today.
POST vs other Methods: The other MS folks were surprised about the
objection. They said that they thought all current servers were capable of
extra methods (certainly NS and Apache). They had not heard the objection
from Novell in the DAV group so they had assumed that the NW web server was
equally capable.
Why use non_POST: Firewalls was the main issue - it gives good granularity
for an admin (I explained the group's feeling about firewall debates). There
are other reasons for it which we did not have time to cover. We will
present the reasons for and against both the XML and different method usage
in Hawaii.
Hope this all helps
Paul Moore