Re: Report on testing of the link relations registry

Hi Silvia,

thanks for the feedback!

On 17.08.2010 13:36, Silvia Pfeiffer wrote:
> Indeed I am looking at this as a curious onlooker, so might not have all
> the information, apologies if I am jumping to conclusions.
>
> Actually, I was not aware that the IANA registry was only just now
> created, but assumed it had existed for a while. Seeing as it's new, it
> should be even easier to fix it with some cooperation from both sides.

The specification that defines the registry is already improved and 
can't be easily be changed before publication as RFC (editorial fixes 
and limited clarifications may be possible).

It has been discussed over several years, and we were careful to address 
the comments from over here; for instance the licensing, the machine 
readability and the metadata handling were all added based on requests 
from over here.

> The problems that I can see - from my limited reading of just this thread:
>
> * If the registry does not solve the problems that the people have that
> want to register a new link relation, then it's not appropriate. What I
> read was that only some of the link relations were acceptable for the
> current registration process, others were rejected as not appropriate. I
> read that some link relations were rejected because they are only
> appropriate for HTML. That is not a good position to be in and not a

Could you clarify which one you are referring to? I don't recall 
anything has been rejected as being "HTML only".

> forward-looking position. If it is required to make notes on the link
> relations what technologies they apply to, then that's a field that the
> registry needs to add.

Yes. But on the other hand it would be good to have a solid consensus 
behind having that distinction. Right now, HTML5 disallows a few link 
relations on <link>, and as far as I can tell, that doesn't make any 
sense at all.

> * Also, I read that the registration process seemed to be non-obvious.
> An online form where the respective fields that need to be provide are
> listed, together with information boxes (e.g. to explain what type of
> specification is required - would a wiki page somewhere suffice or is a
> W3C recommendation or an RFC required). Incidentally. I've seen RFCs
> where a link to a wiki page that was in control of a reliable
> organisation was acceptable.

Changing the actual requirements would require publishing an update to 
the spec. It has been approved the way it is with IETF consensus; and it 
most certainly was discussed over here before we got there.

Simplifying the procedure for newcomers actually has been discussed; it 
wouldn't change what the spec says, but would provide tools to make 
things simpler. We're not there yet.

> Other than that I didn't really see too much of an issue. Anyone who's
> ever registered an RFC or a new mime type knows that there are several
> hoops to jump and that's just what it takes, as long as they are not
> impossible to achieve.

Yup.

Best regards, Julian

Received on Tuesday, 17 August 2010 11:51:13 UTC