Re: I-D ACTION:draft-nottingham-http-link-header-01.txt

On 18/03/2008, at 8:38 PM, Frank Ellermann wrote:

> Mark Nottingham wrote:
>
>> If there are a few nodding heads here
>
> Nod, in essence.  Some nits and questions:
>
> Maybe I missed it, but please mention that <URI-reference>
> is specified in STD 66.  Some folks confuse URIs with IRIs,
> and then get it completely wrong.  This confusion could be
> bad enough for a security consideration.

Ack.

> You mention "Made", but you don't list it in the registry.
> IMO it should be "made", and you should mention that such
> relative URIs are supposed to be case-insensitive.  Maybe
> the IANA Webmaster will be annoyed by this idea, and this
> needs work (?)

Will take a look.

> Maybe add "updates 2626" (?) to get rid of it for 2616bis
> Maybe deprecate "rev" (?) as proposed by the HTML5 cabale

Yeah, been thinking about that. What do others think? Would anyone  
shed a tear if we didn't support rev? This would effectively make it a  
HTML(<5)-specific link-extension.

> The XLink stuff muddies the water, please remove it.

*shrug* I thought it might come up, but AFAICT XLink *is* dead.

> We need to find out where "first", "last", and "payment"
> were defined.  Has <http://www.w3.org/TR/relations.html>
> a status that can be referenced in a RFC ?

I registered the Atom 'first' and 'last'; I'll dig up the e-mail. I  
can chase the definition of 'payment', hopefully.

One substantive question for reviewers: is the registration policy too  
open?

Cheers,



--
Mark Nottingham     http://www.mnot.net/

Received on Wednesday, 19 March 2008 01:07:26 UTC