Re: Report on testing of the link relations registry

Henri Sivonen, Thu, 2 Sep 2010 16:20:43 +0300:
> On Sep 2, 2010, at 01:03, David Singer wrote:
> 
>> IANA is very successful, respected, and useful, for many many types.
> 
> What Web-related registrations is IANA successful and useful for? Off 
> the top of my head, the language tag registry is pretty complete. 
> Other than that, the MIME registry doesn't usefully and successfully 
> match reality (e.g. image/svg+xml), the charset registry doesn't 
> match reality (see additional aliases in HTML5) and the URL scheme 
> and HTTP header registries are totally out of sync with the deployed 
> practice.

It may be worthwhile to compare the successful language registry with 
the others, trying to find out why one works quite well, while some 
others does not.

Some thoughts about why the language registry works well:  its 
community is very dedicated, it seems to me, with people who take their 
responsibility deadly seriously. It is also a registry where  it is 
relatively easy to get new registrations into the registry. It is also 
a registry with detailed rules, including rules for 
deprecation/grandfathering (Ian doesn't take those takes deprecation 
seriously enough, as documented in ISSUE-118 [1]). May be it also 
matters that language is a field where internet engineers are not the 
ultimate experts and where this fact is general accepted by those who 
maintain the registry  (in comparison, consider how difference between 
'text/html' and 'application/xhtml+xml' has become completely mythical, 
with claims in all directions w.r.t. to who if anyone understands the 
issue at all ). The language tag registry also has tight cooperation 
with other, related registries and authorities.
-- 
leif halvard silli

Received on Thursday, 2 September 2010 15:02:01 UTC