On Wed, Jul 8, 2009 at 5:41 PM, Sergey Chernyshevsergey.chernyshev@gmail.com wrote:
I see your point - video is clearly more popular then RDFa and if you're willing to go off-standard to support it, it's might be a reasonable decision for a site like Wikipedia. Not sure what is the rush for that and why can't it wait till HTML 5 spec becomes a recommendation.
The HTML 5 FAQ has useful info on this:
http://wiki.whatwg.org/wiki/FAQ
See especially "When will HTML 5 be finished?" and "When will we be able to start using these new features?" HTML 5 likely will not reach even *Candidate* Recommendation stage until 2012, and might take until 2020 or later to get to Recommendation status. It's a very large spec, and there's absolutely no reason not to use the parts that are fully fleshed out, implemented, and stable just because some other parts are less stable. As I said before, we've always done this with CSS; and this is the official position of the ones responsible for writing and implementing the HTML 5 specification.
I'm not that familiar with HTML 5 support in modern browsers to state that there are going to be regressions with some other things, but it might be another thing to consider, although Wikipedia might be big enough to be a driving force in such decisions.
We are talking about using only polished, finalized features that are implemented in stable browsers which have undergone considerable testing. Unless you can point out specific problems that might arise, there's no reason to anticipate more risk of unexpected problems with HTML 5 than with any other new type of functionality we deploy. There might hypothetically be some useful things that work in XHTML 1 but not HTML 5, but there are *definitely* a number of useful things that work in HTML 5 but not XHTML 1, which have already been outlined in this thread. (And in practice, the WHATWG has made a point of incorporating all unequivocally useful features from XHTML in some form.)
Keep in mind that changing MediaWiki to output valid HTML 5 (modulo GIGO) instead of XHTML 1 on a normal page view would probably take under 20 lines of code changes. I could do it in five minutes. This is an *extremely* small change. Each specific feature of HTML 5 that we use after that can be evaluated for deployment on a case-by-case basis, just as we'd evaluate any other new technology like web fonts or RDFa. If problems do arise from switching to HTML 5 per se, it would be easy to change back to XHTML.