W3C home > Mailing lists > Public > ietf-http-wg-old@w3.org > January to April 1996

Re: Byte ranges & Document title (question)

From: Glenn Adams <glenn@stonehand.com>
Date: Thu, 18 Jan 96 12:09:51 -0500
Message-Id: <9601181709.AA19487@trubetzkoy.stonehand.com>
To: mirsad.todorovac@fer.hr
Cc: http-wg%cuckoo.hpl.hp.com@hplb.hpl.hp.com

    From: Mirsad Todorovac <tm@rasips2.rasip.etf.hr>
    Date: Thu, 18 Jan 1996 17:23:45 +0100 (MET)

    Is there any hope that there could be at least a particular
    solution, instead of saying "It shouldn't be done at all, because
    in complex documents it becomes infeasible" ?

I don't think I said it "shouldn't" or that it was "infeasible".  More
precisely, I should say that byte range fetches of HTML (and SGML) are
insufficient for the purpose of parsing and/or displaying Web documents.

On the other hand, it is possible to reliably process fragments of
HTML as long as sufficient "fragment context information" is provided.
This information turns out to be a subset of the information needed
to process an arbitrary SGML fragment (where a fragment is either a
single element or a contiguous sequence of elements).

SGML Open, an organization consisting of primarily SGML vendors and
users, is developing a technical resolution (TR9601:1996) which deals
with this very issue.  A white paper and a committe draft of this
resolution were reviewed at the recent SGML Open Tech. Committee
meetings that followed SGML '95 in Boston in early December.

If one combines a fragment context description along with a byte range
delimiting the fragment, then parsing/displaying the fragment on its
own is quite feasible.

Regards,
Glenn Adams
Received on Thursday, 18 January 1996 09:16:53 EST

This archive was generated by hypermail pre-2.1.9 : Wednesday, 24 September 2003 06:31:43 EDT