|>> Dave Raggett writes:
|>>
|>> > I believe that the time has now come to develop HTTP as an IETF backed
|>> > protocol with the standard defined as a sequence of RFCs.
|>
|>It sounds like the MIT/CERN consortium (with Tim Berners-Lee) has the same
|>goal, of enhancing HTTP in a controlled manner, and as an open standard free
|>from proprietary commercial interests. What would be the relationship between
|>the proposed IETF working group for HTTP (in the above message), and the
|>MIT/CERN efforts? We would want to avoid the Unix syndrome of multiple
|>standards. Maybe the two groups could work together to propose and ratify the
|>same extensions to the HTTP standard, with the IETF body giving the unified
|>standard enhanced credibility?
Tim is unlikely to respond so I will.
First off Dave did contact me and a number of CERN WWW developers before
launching this. While it is diffult to talk about an organisation that does
not exist (yet) Dave is one of the few people besides Tim who can make
statements likely to be identifiable with future W30 policy with a significant
probability of a high correlation factor between prediction and practice.
I hope that is vague enough for everyone to understand.
On the subject of W3O it is not Tim's or anyone else involved in W3O to go
round screwing other peoples standards, pinching their lunches, atarting
turf wars etc. W3O is an organisation to do something of an X11 with a
distinct IETF flavour. Many of the people involved in W3O will be IETF
stalwarts. Many will be contributors to www-talk. W3O is needed because
unlike other IETF protocols we have the significant disadvantage of a
very highly compressed time frame. We have commercial interests to satisfy,
hardware vendors wanting to get generic standards and willing to chip in
with personel etc and governments wanting to spend money on IT.
There are already severe other constraints on us. The HTML spec is quite
rightly having to go through lots of hoops to meet the needs of the ICAD
(disabled & special needs) people. Currently the maths end of HTML is
seeing quite a lot of improvement in other areas as a result. Legitimate
constraints are often useful. It is the illegitimate ones that pop up because
of half cocked notions of optimisation that are often the most dangerous
ones.
At some point I would see the security side of things splitting into a separate
working group anyway. I don't think we should suggest a separate group quite
yet though, it really is a question of overlap between PEM and HTTP. Perhaps
a joint subcommittee might be a usefull adjunct to try to sort out the
interface.
On the Chicago front I was hoping for a Pizza of a feather session one evening
on the subject of http. If a BOF is booked please don't clash with the
security or HTML+ sessions since certain flag carriers have to attend both
who might just have a few things to say on http as well :-)
Last time HTTP was discussed in the specialised servers and gateways
workshop because there was so much overlap.
-- Phillip M. Hallam-BakerNot Speaking for anyone else.