- From: Leif Halvard Silli <xn--mlform-iua@xn--mlform-iua.no>
- Date: Sun, 21 Feb 2010 22:05:03 +0100
- To: "Tab Atkins Jr." <jackalmage@gmail.com>
- Cc: Julian Reschke <julian.reschke@gmx.de>, tantek@cs.stanford.edu, Krzysztof MaczyĆski <1981km@gmail.com>, Toby Inkster <tai@g5n.co.uk>, public-html@w3.org
Tab Atkins Jr., Sun, 21 Feb 2010 09:26:46 -0600: > 2010/2/21 Leif Halvard Silli <xn--mlform-iua@xn--mlform-iua.no>: >> Julian Reschke, Sun, 21 Feb 2010 09:52:27 +0100: >>> On 21.02.2010 09:42, Tantek Celik wrote: >>>> In short, nothing for meta name (invisible data antipattern), and >>>> nothing for data-* which is not intended for use as shared >>>> vocabulary (as far as I can tell from reading the spec). >>> >>> ...which *does* mean that meta/@name needs to be included. >>> >>> I understand the antipattern argument, but that's orthogonal to the >>> definition of @profile. >>> >>> I agree with the part about data-*. >> >> So then we are at least two persons that think data-* *should* be >> included. > > If I'm reading Julian correctly, then no, he thinks that @profile > should not affect data-* attributes. (Sorry, meant Toby.) > Data-* interactions with @profile is not required for any existing > use-cases, and it is explicitly anti-recommended in the current HTML > draft as data-* attributes should only carry private semantics, not > publicly-shareable ones. I don't see why the spec's space ship element example, with its related data-* attributes, _has_ to be private (whatever "private" means). That example works exactly like I imagine: A certain class name together with a specific element, represents the "space ship element". And the "space ship element" in turn has some required/optional data-* attributes. > If you believe you have concrete use-cases > that show the contrary, please indicate them on the page and in the > manner that Tantek requested. That will help ensure that maximum > consideration is given to the idea. Will add it (unless someone comes before me.) We might not be as much in disagreement that it seems though: I also think that ARIA/@role belongs together with @profile. Not that a profile can be used to invent aria values. But @profile can be used to specify classes - aka "elements" - that in turn requires/recommends specific aria values. The same goes for data-*: It might not be a good idea to specify that the presence of a particular profile means that data-myattr="" has a certain meaning. But it is IMHO another thing to say that @data-myattr="" has a specific meaning inside <span profile="http://my.example.com" class="my-class" data-myattr="value">. So, yes, @data-* should be in scope, but only subsequently. -- leif halvard silli
Received on Sunday, 21 February 2010 21:05:38 UTC