- From: Piotr Jurkiewicz <pjurkiew@agh.edu.pl>
- Date: Thu, 7 Apr 2016 03:20:30 +0200
- To: ietf-http-wg@w3.org
(I sent this comment already two weeks ago to ietf@ietf.org, as
suggested in the Last Call announcement, but received no response or any
other kind of acknowledgement. Therefore, I am sending this to this
mailing list, hoping to find someone interested in it here.)
This document requests a registration in the HTTP Content Coding
Registry. Its authors proposed the 'br' token, which is an abbreviation
of the algorithm name ('brotli').
However, by reviewing existing entries in the HTTP Content Coding
Registry, you can notice that so far established convention is to use
*full* algorithm name, rather than its abbreviation. For example:
- 'gzip' is used instead of 'gz'
- 'compress' is used instead of 'z'
- etc.
This convention is also respected by implementations when they create
unofficial/non-standardized tokens. For example lighttpd and lynx
support bzip2 compression and use 'bzip2' token for indicating it,
rather than 'bz2' abbreviation.
I would urge to not break this convention and maintain the existing
consistency, that is, to select and register 'brotli' as a HTTP Content
Coding token for Brotli Compressed Data Format, instead the 'br'
abbreviation.
Piotr Jurkiewicz
Department of Telecommunications
AGH University of Science and Technology
Kraków, Poland
Received on Thursday, 7 April 2016 01:20:58 UTC