Re: HTML Data Guide draft

Thanks Jeni,

One other issue that probably needs to be addressed here, and I think is an open RDFa and TAG issue, is URI conversion of @href, @src, @data and @about. I believe you created a TAG issue to see if this should be a bug against HTML [1]. The discussion seems to have been side-tracked in issues other than strictly conforming IRIs. The issue for this document, and an open issue for RDFa, is should the RDFa spec provide a warning about such  IRI conversion, and recommend the use of @about or @resource to avoid such conversion, this could also go in this document.

Another alternative discussed last week on the RDFWA WG telecon was if we should perform %-decoding of @href, etc. This would "repair the damage" done by the transformation, but we would be out of sync with other RDF serializations  (such as Turtle) which AFAIK do not require %-decoding (they do, however, do other decoding of escaped characters).

The best outcome would be if a valid IRI used within HTML5 attributes was not converted to a URL.

Anyway, perhaps you can provide an update about where this is going and make an appropriate note in the Data Guide.

Gregg

[1] URL/URI/IRI resolution in HTML5/RDF/RDFa: http://lists.w3.org/Archives/Public/www-tag/2011Nov/0003.html
On Dec 12, 2011, at 9:06 AM, Jeni Tennison wrote:

> Thanks Gregg,
> 
> On 11 Dec 2011, at 18:37, Gregg Kellogg wrote:
>> Looks good Jeni. Do we want to mention Turtle in HTML at all?
> 
> I added this in a Scope section:
> 
>  https://dvcs.w3.org/hg/htmldata/raw-file/default/html-data-guide/index.html#scope
> 
>> Regarding <time> element value range in RDFa. Note that the current resolution of the WG [1] is to accept datatype ranges described in the HTML5 document where they match XSD; this would go in the html+rdfa spec, which is on a different publication timeline than rdfa-core and xhtml+rdfa. Presently, this includes xsd:date, xsd:dateTime, xsd:time, xsd:duration, xsd:gYear, and xsd:gYearMonth. The value can be expressed either using the @datetime attribute or the element's value, but must be a complete lexical match, including possible whitespace, otherwise a plain literal is generated.
> 
> I *think* that this is covered in:
> 
>  https://dvcs.w3.org/hg/htmldata/raw-file/default/html-data-guide/index.html#dates-and-times
> 
> (which I expanded following comments) but let me know if not.
> 
>> The use of @resource within RDFa Lite is still under discussion. My view is that it should be added, and possibly replace @about. Manu feels that @about was long considered and conveys more inherent semantic meaning than @resource. The current draft of RDFa Lite 1.1 discusses this and solicits feedback on the topic.
> 
> Yep, I know. Have you got some suggested rewording for:
> 
>  https://dvcs.w3.org/hg/htmldata/raw-file/default/html-data-guide/index.html#microdata-and-rdfa-equivalencies
> 
> ?
> 
>> RDFa Core 1.1 and HTML+RDFa define just three link relations: describedby, license and role. XHTML+RDFa has the existing step.
> 
> I've rejigged
> 
>  https://dvcs.w3.org/hg/htmldata/raw-file/default/html-data-guide/index.html#link-relations
> 
> Let me know if that's not accurate.
> 
>> One thing I would suggest as a best practice for authors is that the encorporate a data round-tripping step when testing, to ensure that the data extracted from the page matches the data used to create it. For example, if data is modeled using Turtle, this can be used through a template-based writer to create the HTML markup. This should then be distilled back to RDF to ensure that the two reasonably match. The same process can be done for JSON-based representations.
> 
> 
> I've added that suggestion at:
> 
>  https://dvcs.w3.org/hg/htmldata/raw-file/default/html-data-guide/index.html#testing
> 
> Thanks!
> 
> Jeni
> -- 
> Jeni Tennison
> http://www.jenitennison.com
>

Received on Tuesday, 13 December 2011 02:58:30 UTC