Re: Question about philosophy

On Mon, 2015-04-13 at 19:05 -0400, Matt Garrish wrote:

> There’s a difference between what @role allows and what validators > 
allow. There isn’t a restriction on the values you can use per the > 
specification, but HTML validators restrict the attribute to the set > 
of roles defined in the ARIA 1.0 specification. Unrecognized > 
semantics are ignored and AT either fallback to a recognized value > 
in the attribute or to the implied semantics of the element, but you > 
may still get the appearance of non-compliance.
>
> It would be nice if unrecognized unprefixed values were errors and > 
unrecognized prefixed ones not, but that’s not a determination we > 
can make here, and not defined or required by ARIA.

No, but it's feedback we can give to the PF Working Group.

We're here to change the Web :-)


>
> At this point, though, we also fall into the problem of labels as > 
prefixes. There is nothing to say who owns “dpub-“; it lacks a > 
unique identifier. So slapping a “prefix” on your vocabulary isn’t > 
even completely safe short of some additional extension modifier, or > 
using RDFa prefixes.

The ARIA spec could have a built-in list of prefixes, too, rather like 
the data-* attributes in HTML 5, where the prefix "data-" is pre-
registered.

> [...]

> It’s easy to say all abstracts are summaries, but for whom is that > 
helpful? It’s less information for the reader. It’s less specificity > 
for the publisher. It’s problematic across domains leading to > 
specialization.

It's useful to anyone processing a document who understand something 
about the context and domain.

This is the same argument we had about XML on the Web -- what use is a 
Florbl element if no-oneknows what it means? Well, usually the 
Florbelists know. But how do other people find out? And how does a 
search engine display a Florblin in a result snippet?

So it might be that some form of subclassing approach would work --
  role="abstract.wsj"


Liam

Received on Tuesday, 14 April 2015 05:16:42 UTC