W3C home > Mailing lists > Public > ietf-http-wg@w3.org > April to June 2008

Re: Content-Disposition (new issue?)

From: Julian Reschke <julian.reschke@gmx.de>
Date: Fri, 20 Jun 2008 14:38:36 +0200
Message-ID: <485BA4CC.7060108@gmx.de>
To: Frank Ellermann <hmdmhdfmhdjmzdtjmzdtzktdkztdjz@gmail.com>
CC: ietf-http-wg@w3.org

Frank Ellermann wrote:
> ...
>> Well, we need one encoding that can represent all Unicode code
>> points, and UTF-8 comes to mind for that.
> 
> Sure, maybe I don't understand what you are proposing.  The 2231
> example is:
> 
> | Content-Type: application/x-stuff;
> |  title*=us-ascii'en-us'This%20is%20%2A%2A%2Afun%2A%2A%2A
> 
> This would also work for utf-8, iso-8859-1, and windows-1252 --
> these charsets use the same encoding for u+0020, u+002A, "This",
> and "fun".

That's true. But what we need is something that also works for "߀" & 
friends *reliably*.

> I thought that you want to limit the choice to utf-8, which is
> kind of odd in an environment with "default" iso-8859-1.  Maybe
> somebody has implemented RFC 2231 in the last decade for HTTP.
> (without those odd continuations relevant for mail, hopefully)

FF and Opera support RFC2231, and they do support UTF-8.

Reminding UAs to implement RFC2231 without stating that they really need 
also support the UTF-8 encoding seems like a bad idea.

And as there's no way for feature discovery/negotiation, sites really 
need a universal encoding they can count on.

>> Support for RFC2231 in UAs would be totally pointless if we
>> couldn't require recipients to understand it. So why produce
>> anything else?
> 
> How can an UA not understand iso-8859-1 if it is the "default",
> or not support us-ascii as proper subset of utf-8, or not grok
> windows-1252, which is what folks anyway do when they allegedly
> use iso-8859-1 ?  You could say "MUST at least support us-ascii
> and utf-8, and using one of these charsets is RECOMMENDED", or 
> similar.

Not sure what mentioning "us-ascii" buys as here, as the UTF-8 encoding 
is the same anyway. So requiring *both* doesn't give us any additional 
functionality whatsoever.

The second part of the sentence sounds ok, but seems to be really 
useless. If senders can rely on recipients understanding UTF-8, there is 
zero reason to ever try something else.

BR, Julian
Received on Friday, 20 June 2008 12:39:20 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 27 April 2012 06:50:48 GMT