Re: [link-relations] [apps-discuss] Fwd: I-D Action: draft-ohye-canonical-link-relation-01.txt

[Mykyta Yevstifeyev <[e-mail]>]

12.09.2011 15:59, Julian Reschke wrote:
> On 2011-09-12 14:55, Mykyta Yevstifeyev wrote:
>> Maile,
>>
>> Could you please justify why do you want to constrain the requirements
>> of your document with HTTP model?
>>
>> Mykyta Yevstifeyev
>> ...
>
> I believe that while the link relation itself is not restricted to 
> specific protocols, it still makes sense to explain how it should be 
> used with HTTP URIs.

There are plenty of other URI schemes, beyond 'http' and 'https', and I 
don't see reasons we may overlook them defining semantics for these two 
only.  Otherwise, the authors should have "The 'canonical' link relation 
type for 'http' and 'https' URIs" rather than "The 'canonical' link 
relation type" in the title.

BTW, there already is a -02 version: 
http://tools.ietf.org/html/draft-ohye-canonical-link-relation-02.  
However, there are still the two places where some requirements are 
constrained by HTTP protocol:

>     o  A URI that serves a 4xx error code (Section 10.4 of [RFC2616]).

and

>     2.  Permanent HTTP redirects (Section 10.3.2 of [RFC2616]), the
>         traditional strong indicator that a URI's content has been
>         permanently moved, could not be implemented in place of the
>         canonical link relation.

And I still think this HTTP-centric approach is unacceptable.

Mykyta Yevstifeyev

>
> Best regards, Julian
>