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

On May 11, 2009, at 1:34 AM, Sam Ruby wrote:

> Maciej Stachowiak wrote:
>> Hey Ben,
>> 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. You also described  
>> most of the deployments of RDFa in text/html you cite below as  
>> "experimental", and said they should not be considered to make the  
>> text/html syntax a done deal.
>> Given this, I do not give great weight to your objection. You can't  
>> have it both ways.
> There is the (considerably smaller) issue of what markup should be  
> valid in XHTML5.

True, and conflicting use of the name "property" is definitely  
problematic in this context.

> It also appears that Ben's objection could be addressed by simply  
> picking a name other than 'property' for that particular attribute.   
> To date, HTML5 development has given consideration not only to specs  
> as written, but also to tooling as deployed.  I'd go further and say  
> that HTML5 has gone to great lengths to NOT break even small  
> percentages of the web as it is practiced today, independent of  
> whether or not such practices are well advised or ill advised.

Perhaps Ben can clarify his objection, but my understanding is that  
the use of the name 'property' is not the only detail he objects to.  
It seems that his primary objection is that the proposal "gratuitously  
ignores much existing spec work and much existing deployment," with  
the conflict in attributes named "property" an additional concern.

> 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".  I would hope to  
> see a similar level of deployment effort for this proposal before we  
> make a final determination as to whether or not it makes the cut for  
> Last Call.

I don't have strong feelings about Ian's proposal either way. I have  
given it some brief study, but I agree that it's a late-breaking  
feature not yet proven by implementation or use. I would not have a  
problem with HTML5 going to Last Call without it.

However, I do think that Ben's broader objection is ill-founded.

I think a fruitful exercise to apply to this proposal would be to  
determine wether microdata markup following this proposal could  
readily be extracted as RDF. RDF already has many concrete syntaxes; I  
gather that what RDF experts consider important is the underlying data  
model. Can data marked up as Ian suggests be extracted as RDF triples?  
It seems to me this is what is most important for those who wish to  
interoperate with an RDF toolchain.


>> Regads,
>> Maciej
> - Sam Ruby
>> On May 10, 2009, at 4:38 PM, Ben Adida wrote:
>>> Sam Ruby wrote:
>>>> It appears that Ian is on the cusp of making a proposal.  It may  
>>>> turn out to be something that people can live with, and if so,  
>>>> I'll be glad to declare consensus
>>> The proposal is up, and, as Creative Commons rep, I cannot live  
>>> with it (it's not even close, frankly.)
>>> First, this gratuitously ignores much existing spec work and much  
>>> existing deployment (Yahoo, CC, MySpace, Slideshare, the UK  
>>> government, the US government, etc.) with a number of use cases  
>>> that are simply not taken into account (Manu has discussed these  
>>> at length on the WHATWG list). When another spec solves the  
>>> problem and has been deployed by significant players, the first  
>>> step is to consider how that spec can be integrated to the fullest  
>>> extent.
>>> So, I cannot live with something that throws away existing  
>>> important implementations of the *exact* same use cases for no  
>>> valid technical reason. The cost to existing implementors is far  
>>> too high.
>>> In addition, this proposal *specifically* conflicts with RDFa by  
>>> reusing RDFa attributes (i.e. @property) with a different  
>>> interpretation. In other words, of all possible approaches to the  
>>> problem, the HTML5 group chose an approach that specifically  
>>> conflicts with the only other existing W3C spec for the given use  
>>> cases. I think this may be a W3C first.
>>> I absolutely cannot live with that.
>>> I note, as a side point, that it's fairly clear this conflict was  
>>> by design (since it was said that @property is "borrowed from  
>>> RDFa"). In other words, whereas typical W3C groups go out of their  
>>> way to prevent conflict with other specs, this group is currently  
>>> actively creating conflict.
>>> -Ben

Received on Monday, 11 May 2009 08:56:45 UTC