W3C home > Mailing lists > Public > ietf-http-wg-old@w3.org > May to August 1997

Re: Rewrite of feature tag syntax rules

From: Roy T. Fielding <fielding@kiwi.ICS.UCI.EDU>
Date: Thu, 15 May 1997 16:09:28 -0700
To: http-wg@cuckoo.hpl.hp.com
Message-Id: <9705151615.aa20944@paris.ics.uci.edu>
It seems to me that the whole idea has gotten out of control.
There simply is no point in exchanging feature sets over the
wire if they can't be exchanged efficiently, and the draft 00
syntax of tags was already too big.  Hell, I'd have a hard time
convincing people to use pure bit vectors -- the number of possible
features is simply too big.

It makes far more sense to me to do this stuff by reference to
well-known sets using a relative URI to a common base, e.g.

    iana:/http/featuresets/

and then populate it with common sets, e.g.,

    ht2 ht3.2 ht4 cougar ns1.1 shkwv

with each set consisting of feature_name/range pairs and an entry
indicating what other sets are also covered by this set [inclusion
is necessary in order to avoid huge numbers of set names being exchanged
on the wire].  Features would then be a simple list of URIs, and the
authors can use non-relative URLs if they want to screw themselves.
Optimize for the common case.

....Roy  [and pity the fool who says there is no need for relative URNs]
Received on Thursday, 15 May 1997 17:02:30 EDT

This archive was generated by hypermail pre-2.1.9 : Wednesday, 24 September 2003 06:32:42 EDT