- From: Jeffrey Mogul <mogul@pa.dec.com>
- Date: Mon, 05 Jan 98 10:21:43 PST
- To: http-wg%cuckoo.hpl.hp.com@hplb.hpl.hp.com, mogul@pa.dec.com
Fred Douglis wrote, in reply to Simon Barratt ( Barog ):
>What discussions of compression has their been? Will a simple compression
>algorithm be implemented into HTTP (only as an option, so that slower
>clients can choose not to retrieve compressed data).. The data could be
>compressed already, and sent out in this form... Also an uncompressed
>version could still be stored to send to the slower clients (who couldn't
>decompress realtime).
There were two papers in SIGCOMM'97 that discussed the benefits of
compressing HTTP data. An internet draft to describe protocol
support for both compression and delta-encoding (a form of
compression that uses state from a previous version of a resource)
is in preparation. Jeff Mogul, of DECWRL, is the primary
"instigator" on this.
Clarifications:
Actually, to answer Simon's question: HTTP/1.1 already does support
optional compression at several phases of the interaction. I believe
that it provides all of what Simon was asking about. (This is true
of the most recent Internet-Draft on HTTP/1.1; RFC2068 has some minor
bugs in its support for compression.)
Therefore, the Internet-draft that I am in the process of writing
(please do NOT ask me for a copy yet!) on delta-encoding says almost
nothing about compression, because this is a (mostly) solved problem.
-Jeff
Received on Monday, 5 January 1998 11:12:40 UTC