Re: PEP Battle Plan [rexmit, garbled]

At 05:34 PM 10/19/96 -0400, Rich Salz wrote:
>>> Progress on an extension mechanism is essential because it is the
future of
>>> 1.x and binary encodings of it.
>>
>>I realize this is heresy here, but I have to wonder if it's worth
>>building the extension mechanism into HTTP.

Every protocol needs an extension model, otherwise they will become
obsolute before you know it. What PEP is all about is realizing that the
current RFC-822 extension mechanism inherited by HTTP in many cases isn't
enough. The PEP framework provides three types of services:

A) PEP gives the parties the possibility of enquering and enumerating
available extensions.

B) PEP gives you the possibility of defining three important attributes of
any extension:

	- consequence - what is the consequence of having / not having an extension? 
	- ordering - does extension A come before or after B?
	- scope - this is partially solved with the HTTP/1.1 Connection header but
it needs to be bound to the PEP frame work as well

C) PEP helps avoiding header name collisions.

Neither of these services are provided by the traditional model of simply
adding new headers. I don't say that the existing model isn't enough in
some cases - just not en all of them.

>I don't think there's anything heretical about "declare victory and
>move on."
>
>I, ,too, don't think PEP is the future of HTTP.

Hang on, after much discussion at the Montreal IETF http-wg meeting, it was
decided to continue working on Content negotiation, User Agent, and PEP.
Please check the minutes as posted to the mailing list

	http://www.w3.org/pub/WWW/Protocols/IETF/9606-Montreal/http-wg-minutes.txt

This has nothing to do with victory or any other terms borrowed from the
battlefield. If you have constructive critisism of the current draft as
issued August 19th then you should forward them to this list.


--
Henrik Frystyk Nielsen, <frystyk@w3.org>
World Wide Web Consortium, MIT/LCS NE43-356
545 Technology Square, Cambridge MA 02139, USA

Received on Sunday, 20 October 1996 15:07:36 UTC