[whatwg] Micro-data/Microformats/RDFa Interoperability Requirement

[Ian Hickson]

On Sat, 9 May 2009, Manu Sporny wrote:
> > 
> > For rel-license, the HTML5 spec defines the value to apply to the 
> > content and not the page as a whole. This is a recent change to match 
> > actual practice and I will be posting about this shortly.
> 
> Hmm, yes - after re-reading the definitions, they do differ... 
> especially in how the hAudio Microformat uses rel="license". I find the 
> HTML5 one to be very problematic. Microformats rel="license" is better, 
> and the RDFa use of rel="license" is even better (I can go into the 
> reasoning if those on the list are curious).

The definition in HTML5 is just a description of how it's used in 
practice. It's not what I'd design if I had a choice. However, there's not 
much point defining something that doesn't match the reality of the 
situation.


> For example, in HTML5, how do you express 20 items on a page, each with 
> separate licenses? How do you differentiate a page that has 3 primary 
> topics, each with a separate license?

You can't, currently. Can you provide a sample of such a page?


> In short - what's the purpose of rel="license" if a machine can't use it 
> to help the person browsing identify important sections of a page?

rel=license as it stands today is useful only (as far as I can tell) for 
people searching the Web for pages whose main content (e.g. the photo on a 
Flickr page, the text of a blog post on a blog post's permalink, etc) is 
licensed under a particular license identified by URI.


> Afterall, it's only machine readable, isn't it? What's the sense in 
> having rel="license" if a machine can't be sure of the section of the 
> page to which it applies?

Generally speaking, search engines are pretty competent at distinguishing 
the page's main content from ancillary content such as style sheets or 
navigation links.


> More importantly, if you see this as an issue, why don't you see the 
> semantic difference between rel="alternate"[3] in HTML4 and 
> rel="alternate"[4] in HTML5 as being an issue? That case is even worse, 
> exactly the same string - entirely different semantics.

HTML5's definition, again, matches what actual implementations and author 
usage is. HTML5 isn't trying to be compatible with HTML4, it's trying to 
be compatible with legacy content.


> > You would have to ask them. I tend not to argue with implementor 
> > feedback. If they tell me they won't do something, I don't tell them 
> > to do it.
> 
> I would expect the primary editor for a web specification to understand 
> the reason that every single one of their implementors refuse to 
> implement a technique that they use elsewhere in their products.

I believe I do, but I wouldn't want to put words in their mouth on such 
controversial topics.


> I'd love to ask eacho of them why it is perceived as difficult and 
> discuss possible solutions. A public e-mail would be best so that we can 
> discuss on this list, but a private e-mail would be fine as well. So 
> please, if your organization has decided to not resolve prefixes in 
> attribute values, please send me an e-mail.
> 
> After two weeks, I'll check back in with the mailing list to report on 
> the number of responses I received and a summary of the reasoning 
> (anonymized, of course) for the benefit of this community.

Did you receive any responses?

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'