[whatwg] RDFa is to structured data, like canvas is to bitmap and SVG is to vector

Shelley Powers shelleyp at burningbird.net
Sun Jan 18 12:04:28 PST 2009

Dan Brickley wrote:
> On 18/1/09 20:07, Henri Sivonen wrote:
>> On Jan 18, 2009, at 20:48, Dan Brickley wrote:
>>> On 18/1/09 19:34, Henri Sivonen wrote:
>>>> On Jan 18, 2009, at 01:32, Shelley Powers wrote:
>>>>> Are you then saying that this will be a showstopper, and there will
>>>>> never be either a workaround or compromise?
>>>> Are the RDFa TF open to compromises that involve changing the XHTML 
>>>> side
>>>> of RDFa not to use attribute whose qualified name has a colon in 
>>>> them to
>>>> achieve DOM Consistency by changing RDFa instead of changing parsing?
>>> I don't believe the RDFa TF are in a position to singlehandedly
>>> rescind a W3C Recommendation, ie.
>>> http://www.w3.org/TR/2008/REC-rdfa-syntax-20081014/.
>>> What they presumably could do is propose new work items within W3C,
>>> which I'd guess would be more likely to be accepted if it had the
>>> active enthusiasm of the core HTML5 team. Am cc:'ing TimBL here who
>>> might have something more to add.
>>> Do you have an alternative design in mind, for expressing the
>>> namespace mappings?
>> The simplest thing is not to have mappings but to put the corresponding
>> absolute URI wherever RDFa uses a CURIE.
> So this would be a kind of "interoperability profile" of RDFa, where 
> certain features approved of by REC-rdfa-syntax-20081014 wouldn't be 
> used in some hypothetical HTML5 RDFa.
> If people can control their urge to use namespace abbreviations, and 
> stick to URIs directly, would this make your DOM-oriented concerns go 
> away?

Took five minutes to make this change in my template. Ran through 
validator.nu. Results:

Doesn't like the content-type. Didn't like profile on head. Having to 
remove the profile attribute in my head element limits usability, but 
I'm not going to  throw myself on the sword for this one.

Doesn't like property, doesn't like about. These are the RDFa attributes 
I'm using. The RDF extractor doesn't care that I used the URIs directly.

Didn't seem to mind SVG, but a value of "none" is a valid value for 

> cheers,
> Dan
> -- 
> http://danbri.org/

More information about the whatwg mailing list