W3C home > Mailing lists > Public > ietf-http-wg-old@w3.org > September to December 1995

Re: Comments on Byte range draft

From: Benjamin Franz <snowhare@netimages.com>
Date: Sun, 12 Nov 1995 07:03:42 -0800 (PST)
To: Gavin Nicol <gtn@ebt.com>
Cc: montulli@mozilla.com, fielding@avron.ICS.UCI.EDU, masinter@parc.xerox.com, ari@netscape.com, john@math.nwu.edu, http-wg%cuckoo.hpl.hp.com@hplb.hpl.hp.com
Message-Id: <Pine.LNX.3.91.951112063700.8492A-100000@ns.viet.net>
On Sun, 12 Nov 1995, Gavin Nicol wrote:

> >You and Larry are looking at this problem with blinders on.
> >There are many more uses for byterange URL's than simply
> >PDF files.  For instance Netscape 2.0 uses byteranges to
> >request parts of files that it didn't get the last time
> >you came to a page. 
> 
> And then you have to reparse the entire document again, and re-render
> it, possibly ignoring the errors caused by the file being
> incomplete. This is great for small pages, but if you try fetching
> small peices of a 5MB document, it makes no sense. 
> 
> Byte ranges are a lazy replacement for a general naming mechanism.

You still have those blinders on. The whole universe of documents is 
not SGML/HTML/PDF/(favorite text markup language with naming mechanism). 
The ability to restart an interrupted transfer is an item that naming 
mechanisms are insufficiently powerful to handle in the general case. 
Byte ranges are not a 'lazy replacement' - they are the only general 
mechanism for restarting interrupted transfers of documents containing 
arbitrary content.

Such as: you just fetched 100K of an inlined 160K GIF/JPG/PNG file but 
interrupted the transfer by following a link before the transfer was 
complete. You then go back to that page. Now rather than spend a minute 
and a half getting the *same* information - you pick up where you left off.

When the content of the interrupted transfer is arbitrary, 
you have to use a recovery mechanism that *ignores* the content. I have 
continually puzzled why in this particular area (resumption of 
interrupted transfers) the Internet implementation has lagged badly behind 
the BBS community - which has had robust recovery mechanisms for interrupted 
transfers for a number of years.

The question is not *if* they should be in the standard, but *how* they 
should be implemented - in the URL or via Yet Another Header. In the 
URL seems to me to have the advantage of not breaking proxies.

Benjamin Franz

"Heavily annoyed by beginining from 0 when D/Ling 2 meg reports that I 
already D/led half of ten minutes ago."
Received on Sunday, 12 November 1995 05:58:48 EST

This archive was generated by hypermail pre-2.1.9 : Wednesday, 24 September 2003 06:31:35 EDT