Re: deflate

At 02:24 PM 3/20/97 -0500, nemo/Joel N. Weber II wrote:
>Are there any HTTP/1.1 servers which support the deflate method?
>I think I have a correct (modulo bugs) client implementation of deflate in
the
>current (alpha) version of E-scape, and I'd like to test it sometime
>before I release 1.0.

You can do most of your testing simply by adding the "deflate" content
encoding on a file - the server doesn't have to deflate it on the fly.
Jigsaw supports this by editing the resource directly and Apache by the
AddEncoding configuration directive.

This is what we did when we did the work on compression in the paper
"Network Performance Effects of HTTP/1.1, CSS1, and PNG" which is available
from

	http://www.w3.org/pub/WWW/Protocols/HTTP/Performance/Pipeline.html

In addition to this we have also done some testing on how HTML
case-canonicalization affects compression. You can find the study at


http://www.w3.org/pub/WWW/Protocols/HTTP/Performance/Compression/HTMLCanon.h
tml

It shows that lower case HTML tags is better than uppercase and that mixed
casing gives the worst result. You can also find a simple "deflater" tool
here.

Also, we have done some testing on how compression tends to have a good
effect on TCP slow start and delayed acknowledgement. You can find this
note at

	http://www.w3.org/pub/WWW/Protocols/HTTP/Performance/Compression/LAN.html

As you might have guessed from the URLs above - we keep a performance
overview at

	http://www.w3.org/pub/WWW/Protocols/HTTP/Performance/
	
Henrik
--
Henrik Frystyk Nielsen, <frystyk@w3.org>
World Wide Web Consortium, MIT/LCS NE43-346
545 Technology Square, Cambridge MA 02139, USA

Received on Thursday, 20 March 1997 12:35:50 UTC