Re: microdata use cases and Getting data out of poorly written Web pages

Maciej writes:
> I seem to recall you said before that RDFa was designed without  
> considering the requirements of text/html, and that it should not be  
> seen as constraining the text/html syntax.

Actually, that's not what I said. I said that we weren't tasked with 
making this work in text/html, but we worked hard to make it compatible 
anyways, in case things evolved.

Sure, HTML5 is not technically constrained by RDFa, although it's basic 
courtesy to not screw someone else's work up if you can help it. Even if 
all you care about is your work and no one else's, you could stand to 
benefit from looking at other people's experiences, see my specific 
comment below.

Sam Ruby writes:
 > All other features in HTML5 have varying degrees of implementation and
 > deployment experience.  At the moment, Ian's proposal appears to be
 > speculative - as in "seems like it would work".

Yes, it's important to validate theories. Henri's theory that xmlns:foo 
would be impossible/difficult to parse correctly in text/html proved to 
be a fairly weak argument in practice (Google, Yahoo, and my Firefox 
bookmarklet do just fine.)

Ian's microdata proposal uses <link> in the body of an HTML page:

http://dev.w3.org/html5/spec/Overview.html#the-link-element

We tried that with RDFa, but during testing, we realized that Firefox 
repositions <link> to within <head> in text/html mode, so we spent weeks 
tweaking RDFa, testing and re-testing. I think this is a prime example 
of the effort we made to be text/html compatible.

Does HTML5 intend to break with existing browser mechanisms for handling 
<link> and <meta>? I thought HTML5 was trying to standardize 
pre-existing error behaviors.

-Ben

Received on Wednesday, 13 May 2009 17:18:59 UTC