Re: review of content type rules by IETF/HTTP community

2007-08-21 22:04:26 +0200 Sander Tekelenburg:
> At 18:23 +0200 UTC, on 2007-08-21, Leif Halvard Silli wrote:
>> 2007-08-21 17:09:58 +0200 Sander Tekelenburg:

>> Currently the extension method doesn't work out of the box. But it should
>> be relatively easy for browsers to start reading charset suffixes.
> 
> FWIW, iCab already does. (And the file name's charset info overrides the
> document's meta http-equiv's charset value.)

Indeed.
 
>> this just a help. If the charset is specified inside the file but not in
>> the filename, then the browsers will use that charset instead - as they
>> allready do.
> 
> I'm less sure now that I understand what exactly you propose. Do I understand
> correctly that you mean that the charset info in the file name overrides both
> @charset and the HTTP Content-Type charset value?

You said what I meant above: «And the file name's charset info overrides the document's meta http-equiv's charset value.» This of course only counts for offline documents. 

> I'll grant you that, despite my aversion of file name extensions[*], it might
> indeed be an option in the sense of providing a new mechanism.

I think so.

> But Ian's 5th  point, raised in
> <http://www.w3.org/mid/Pine.LNX.4.64.0708202003070.8981@dhalsim.dreamhost.com>,
> remains, doesn't it?

Was that the point about «User Bob uninstalls the new browser and goes back to his old one (or 
   worse, from the vendor's point of view, goes to a competing one)»?

If so, then I cannot see at all why it applies. Ian is speaking about putting a file on a «popular site», that is: putting files online. While I am only proposing a to use new/existing mechanism for offline authoring of documents which could be put on that popular site. This mechanism should on the contrary better the situation, in that the author actually got to test how  this "charset mapping" works before putting it online.

Quoting Roy T. Fielding's reply in this thread:
> The only reason that resources don't get assigned correctly is because 
> testing does not reveal the error.
-- 
leif halvard silli

Received on Tuesday, 21 August 2007 21:03:59 UTC