Re: registration request for dns-prefetch, preconnect, prefetch, and prerender link relation types

[Erik Wilde <[e-mail]>]

On 2016-01-30 18:10, Mark Nottingham wrote:
> 5988 doesn't require strong justification for registering a relation type; only that it follows certain rules. So, I'd say the fact that these relations went through an open process is enough to overcome any reluctance to register so many.

that was my gut reaction. seeing 4 or 5 link relations as a "whopping 
amount" seems a bit overstated (after all, the namespace is fairly big, 
so there's no need to be overly conservative trying to conserve space), 
but i guess you mileage may vary.

> WRT 7.4 - the applicable text in 5988 is:
> """
> Registered relation types MUST NOT constrain the media type of the
>     context IRI, and MUST NOT constrain the available representation
>     media types of the target IRI.  However, they can specify the
>     behaviours and properties of the target resource (e.g., allowable
>     HTTP methods, request and response media types that must be
>     supported).
> """
> Putting RFC1918 language in registry entries is a bit weird too.
> So something like this would be clearer:
> """
> Indicates a resource that might be required by the next navigation, and that user agents ought to fetch and execute to improve performance.
> """

i like it. personally. i think that it's a very good idea to not 
constrain link relations any more than they have to be, and as mark 
point out, when it comes to media types, it's not even something you're 
supposed to do.

https://github.com/w3c/resource-hints/pull/54 adopts mark's suggested 
change, and afaict, this does not lose anything in terms of what the 
spec is saying.

thanks and cheers,

dret.

-- 
erik wilde | mailto:[e-mail] |
            | http://dret.net/netdret    |
            | http://twitter.com/dret    |