Re: URI handling for Location

On Tue, 07 Apr 2009 11:16:14 +0200, Anne van Kesteren <>  
> On Tue, 07 Apr 2009 06:18:18 +0200, Mark Nottingham <>  
> wrote:
>> As discussed at the bar bof/informal liaison in SF, this seems to be  
>> more of a URI issue; i.e., HTML and browser vendors need a named  
>> construct ("dirty URI", "hypertext reference", etc.) for what goes into  
>> the process, but from an HTTP perspective, this isn't evident.
>> AIUI the follow-up will happen on the URI mailing list.
> Allowing relative URIs hardly seems a "dirty URI" issue. As for the  
> actual "dirty URI" issue, that isn't solved as long as HTTP keeps  
> insisting the header takes an actual URI rather than such a "dirty URI".

Also, it is not just browsers, this affects all kinds of clients that try  
to do something with Web pages: curl, wget, feed finder clients, etc.

Anne van Kesteren

Received on Tuesday, 7 April 2009 13:09:24 UTC