Re: microformats, profiles, and taking back rel/class names [standardizedFieldValues-51]

On Mon, 16 Jul 2007, Mark Baker wrote:
> On 7/16/07, Harry Halpin <hhalpin@ibiblio.org> wrote:
>> If @profile is lost, and if @rel and @class are given a single
>> centralized repository with *no* way to extend HTML in a principled
>> manner, you are breaking decentralization of the Web, period.
>
> The Web currently uses several registries; DNS, media types, various
> HTTP parameters (methods, response codes, etc..), URI schemes.

*None* of the above options allow you to do what @profile does, i.e.
address the problem of interpretation of attribute values of @class and 
@rel in HTML. Since people are *already* using these attributes in a 
decentralized way (re microformats) and that seems to have no sign of 
stopping, Web Architecture owes these people a way of doing it in
a principled manner  that uses the "Follow-Your-Nose" principle - and 
luckily, HTML 4 and XHTML  1 give us this mechanism. Thus, deprecating 
@profile/@rel in HTML 5 is not only not appropriate, it's broken.

It's very analogous to namespaces - why it's allowed not to use 
namespaces, fundamentally the use of namespaces is slowly increasing and 
to delete namespaces from the next version of XML would be broken.

> I agree it's a bit fuzzy in this case whether centralized or
> decentralized names (or a combination of the two?) is most
> appropriate, but I don't think centralization is prima facie bad.

Some sort of centralization kept to a sensible minimum is almost always 
good, but centralization for  the sake of centralization is bad - 
especially if no other mechanism is provided that gives one that 
functionality, and people are already using that functionality (albeit in 
a broken way with microformats). History is litttered with broken 
centralization schemes like URNs.

  > Mark.
>

-- 
 				--harry

 	Harry Halpin
 	Informatics, University of Edinburgh
         http://www.ibiblio.org/hhalpin

Received on Tuesday, 17 July 2007 00:51:16 UTC