Raw

Sam Ruby points to a nice timbl quote:

HTML is a big community, but there are others communities. Smaller communities are more in need of uri-extensibility than bigger ones.

On the linked write-up of TPAC 2007 - URI-Based Extensibility: Benefits, Deviations, Lessons-Learned, there's a report on Hixie's pov:

Ian Hickson stresses out that there is little value to give URI for microformats for example, and that you still need knowledge of the vocabulary to implement it.

This is more of a technical argument than the social ones put forward there, and hence is a bit easier to deal with. Ok, the details depend a lot of how you interpret implementing a vocabulary, but the implication here, that having a URI for a vocabulary without other knowledge of it is of little value is demonstrably false. 

I won't bother going into the detail of the utility when you use RDF ( how a GRDDL-aware agent can interpret profile-based HTML, how that result can be trivially merged with other data and how a SPARQL engine can allow querying of the combined result, being able to recognise connections across vocabularies - perhaps vocabularies you already do know about). There's a much simpler way of demonstrating the utility of URIs in this context. It follows from the question:

How do you get knowledge of the vocabulary?

With a profile URI you (or your software agent) can follow your nose and GET that knowledge. Otherwise you have know about the vocabulary in advance (or maybe on a system like Google being able to make the connection between a HTML class name like 'fn' and its definition in the vCard RFC). [rel="profile" has been suggested as an alternative, but consider how the meaning of "profile" would be determined...]

Ok, a concrete example - say you don't understand German ( I don't, so I hope this makes sense), but you encounter the following: <span class="de"> Aus Liebe zu Vögeln</span>. How do you interpret that? Without the profile URI, chances are you'll get it wrong.

We tend to rely on a bunch of prior knowledge to implement core specs like HTML itself, but up to a point that's a necessity ( axioms of the www, anyone?). Microformats are well beyond that point. But microformats can easily advertise themselves through the use of profile URIs.

It seems a shame when an intelligent person can summarily dismiss a good idea on the basis that their own (or certain vendors') immediate requirements don't call for it. Worse still when the peer effect makes a group of people act this way. Problem is, once a mind (individual or group) has become closed on an issue for whatever reason, it's not easy to open open it up again.

The reasons put forward around the proposal to drop @profile are at best myopic, at worst self-serving. But what's really irritating about all this is that the folks keen on dropping @profile don't seem to consider the possibility that it might actually be useful to them - if not now, then possibly in the future. HTTP URIs are an extensibility mechanism supported by the Web out of the box. It seems good practice to use it, and encourage the use of URIs generally, irresponsible to minimise the opportunities for using them.

@en