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

Re: NEW ISSUE: content sniffing

From: Barry Leiba <barryleiba@computer.org>
Date: Tue, 31 Mar 2009 21:26:56 -0400
Message-ID: <6c9fcc2a0903311826v1eeaa0a2y795355d155bcdb7@mail.gmail.com>
To: Mark Baker <distobj@acm.org>
Cc: Adam Barth <w3c@adambarth.com>, HTTP Working Group <ietf-http-wg@w3.org>
>> but I think we're better off making the algorithm normative
>> (for those agents that wish to sniff) rather than informative.
>
> I don't know what it means for an algorithm to be normative.  Can you
> elaborate, perhaps by offering the text you'd ideally like to see in
> the HTTPbis spec (assuming the algorithm will be in a separate spec)?

I suspect he's going for something like (abbreviated version here),
"Implementations MAY [or SHOULD NOT] sniff.  If an Implementation does
sniff, it MUST do it this way...."

We should be careful about putting in normative text that we know a
good portion of implementors will ignore.  It's pretty much an
invitation to ignore other bits as well.

I'd be in favour of something like, "The use of different sniffing
algorithms in different implementations creates [these sorts of
problems].  Because of that, settling on common sniffing mechanisms is
important.  To that end, implementations that sniff SHOULD
[...etc...]."  The explanation of why it's important may sway some
implementors, and making it a SHOULD recognizes that some will go
their own way in any case.

Barry
Received on Wednesday, 1 April 2009 01:27:43 GMT

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