- From: Ben Adida <ben@adida.net>
- Date: Wed, 14 Feb 2007 11:36:18 -0500
- To: Simone Onofri <simone.onofri@gmail.com>
- CC: Steven Pemberton <steven.pemberton@cwi.nl>, Ivan Herman <ivan@w3.org>, "Hausenblas, Michael" <michael.hausenblas@joanneum.at>, RDFa <public-rdf-in-xhtml-tf@w3.org>, SWD WG <public-swd-wg@w3.org>, www-qa@w3.org
Simone, Ivan, Steven, I don't want to shut down discussion, but as Chair I have to make an important point: the CLASS vs. ROLE decision is one that we already made as a group, with many discussions and many telecons. If we want to make progress, we need to have some deference for decisions already made unless there is a very good reason to overturn them. So far, I'm not seeing a good reason to overturn a decision, only a reason to bound it (triple bloat). Right now, the discussion is headed in the direction of "let's put everything back on the table." We have to be extremely wary of this approach. It means we have to put a hold on the Primer yet again. It means any question raised sends us back to the drawing board when a small fix might do the trick. After we agreed on CLASS, I spent hours correcting this issue with other groups (who saw my use of ROLE in an earlier Primer draft). Changing it back would be a huge cost again. We must realize that such a change would bear a huge cost, and for what benefit? To address specific points: > So, Using class with semantic meaning, this overload the current use > of @class. It's actually not overloading: Mark has explained very well how CLASS was meant for semantics [1], and then CSS used that as a hook for style. CLASS is not about style, though it can be used for style. > So, if we use on a single page also more dialects we can have: > > ... class="mystile first-dialect second-dialect" ... > > and this is not clear also for humans and create more confusion also > for machines. If we go with namespaced classes, it actually looks quite clear to me. Seems like a matter of taste on this front. > It's a good idea use only "namespaced" class but if we can use a > specific @attribute for semantic layes, this should be the best > choice. New attributes should be used sparingly. HTML already has attributes and elements that provide semantics: LINK, META, REL, REV, and CLASS. Inventing some new attributes is causing us great pain with folks who don't like HTML extensibility. At the very least, we should be minimal in our approach and consider each new attribute carefully. -Ben [1] http://lists.w3.org/Archives/Public/public-rdf-in-xhtml-tf/2006Dec/0018
Received on Wednesday, 14 February 2007 16:36:16 UTC