W3C home > Mailing lists > Public > public-html@w3.org > March 2008

Re: several messages about content sniffing in HTML

From: Philip Taylor <pjt47@cam.ac.uk>
Date: Sat, 01 Mar 2008 01:04:12 +0000
Message-ID: <47C8AB8C.6060700@cam.ac.uk>
To: Julian Reschke <julian.reschke@gmx.de>
CC: Ian Hickson <ian@hixie.ch>, "public-html@w3.org" <public-html@w3.org>

Julian Reschke wrote:
> Ian Hickson wrote:
>> ...
>> [things about defining handling of multiple Content-Type headers]
>> ...
> It would be interesting to know how many pages actually have that 
> problem; chat on IRC indicates it is below 1/1000.

The only data I have is 
<http://philip.html5.org/data/multiple-content-types.txt>, which is too 
limited in scope to see if there are cases where the difference matters 
in practice.

Looking at that data anyway, there were multiple Content-Type headers on 
about 0.04% of those pages (or 0.02% of the hostnames). (For comparison, 
that's about the same number that use <del> or <var>, and about half the 
number that use "Content-Encoding: deflate"). All of them appear to work 
correctly in browsers regardless of the Content-Type processing, since a 
browser that chooses the "text/html" instead of "text/html; charset=..." 
will fall back on some other method to find the right charset.

> BR, Julian

Philip Taylor
Received on Saturday, 1 March 2008 01:04:22 UTC

This archive was generated by hypermail 2.4.0 : Saturday, 9 October 2021 18:44:27 UTC