W3C home > Mailing lists > Public > ietf-http-wg@w3.org > April to June 2009

Re: PROPOSAL: content sniffing [#155]

From: Yves Lafon <ylafon@w3.org>
Date: Thu, 9 Apr 2009 03:35:51 -0400 (EDT)
To: Adrien de Croy <adrien@qbik.com>
cc: Julian Reschke <julian.reschke@gmx.de>, Adam Barth <w3c@adambarth.com>, Mark Nottingham <mnot@mnot.net>, Mark Baker <mark@coactus.com>, =JeffH <Jeff.Hodges@kingsmountain.com>, HTTP Working Group <ietf-http-wg@w3.org>
Message-ID: <Pine.LNX.4.64.0904090322170.3559@ubzre.j3.bet>
On Thu, 9 Apr 2009, Adrien de Croy wrote:

> What about something like:
>
> "When an entity-body is included with a message, attributes of that 
> entity-body are declared in Content-* headers.
>
> Content-Length SHOULD be used to declare the length in bytes of the content 
> where known and where Transfer-Encoding: chunked is not used.
> Content-Type SHOULD be  used to declare the type of the content.
> Content-Encoding MUST be used to declare any encoding where that may have 
> been applied to the content.

Why making a difference betwee CT and CE?
If the content is gzipped, it can be both CT application/gzip or CT 
whatever, CE: gzip. Why should be force CE to be filled if CT is not?

And Content-Length is a weird beast as it applies to both the 
entity-length and transfer-length.
Btw, your last sentence "Content-Encoding MUST be used to declare any 
encoding..." seems to apply more to Transfer Encoding rather than CE, as 
in that case you really want the encoding explicit to present the right 
content once all the connection-level processing is done..

-- 
Baroula que barouleras, au tiéu toujou t'entourneras.

         ~~Yves
Received on Thursday, 9 April 2009 07:36:00 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 27 April 2012 06:51:02 GMT