Change Proposal 005

[Martin Atkins]

Josh Hoyt wrote:
> On 2/9/06, Martin Atkins <mart at degeneration.co.uk> wrote:
> 
>>Mainly my point is that it seems silly to invent a new header when
>>there's a perfectly good one there already — as a de-facto standard, at
>>least — which, as a nice bonus, is also consistant with the way we are
>>referencing YADIS documents within HTML documents.
> 
> 
> But it's not consistent -- the specification uses <meta
> http-equiv="..."> in HTML documents. I think that the new header is
> more elegant, since we expect it (or content-negotiation) to be the
> major pathway taken by implementers. Basically, using the YADIS header
> makes it clear that the HTML path is in there for people who are in a
> limited environment, and that using HTTP to find the XRDS is
> preferred. Using the Link: header makes it seem like HTML is
> preferred.
> 

I don't really follow your reasoning here beyond the fact that YADIS is
using META HTTP-EQUIV.

Nonetheless, I must have missed that change since last I was paying
attention we were using LINK REL like OpenID does.

In which case, I'll back down from the consistency argument, but then I
wonder why we're using META to reference a URL when that's exactly what
LINK is for.

...but never mind. I don't really care about this that much. I just
wanted to put it out there.