Re: Default Vocabulary for datatype: datatypeLibrary attribute

Dominik,

In general, at this point, I would prefer to have a virtual feature freeze in RDFa. By virtual I mean that any new feature should carefully weight the design and implementation costs and should be taken up only if there is a real and documented need for that feature. The reason for this virtual feature freeze is that RDFa is already complex enough (which is one of the problems users raise, rightly or wrongly, v.a.v. RDFa in general) and the goal would rather be a simplification rather than adding new features.

Coming back to your proposal: Adding a such new attribute, though the design is relatively obvious, has to be folded in the specification in a proper fashion. Decisions have to be made on its default values (if any) eg, in an initial context, about its the type that attribute value could take, how it behaves v.a.v. the inheritence in the DOM tree, etc. There is the job of adding its management to the formal processing steps of RDFa which is already fairly convoluted. The corresponding DTD must be changed. Additional tests should be created. Etc.

_None_ of these steps are complicated. But putting them all together is not an insignificant amount of work at this point. On the other hand, as Jenni said, through the usage of the default initial context all XSD datatypes can be used with the xsd: prefix without any declarations. In view of this, my assessment is that costs of introducing that feature significantly outweigh the gains.

Sorry Dominik, but, as a consequence, I am (personally) against this proposal:-(

Cheers

Ivan 



On Nov 22, 2011, at 22:57 , Jeni Tennison wrote:

> Dominik,
> 
> On 22 Nov 2011, at 21:02, Dominik Tomaszuk wrote:
>> We use only one datatype library in @datatype very often. Also in many cases it is just XML Schema Datatypes. I propose @datatypeLibrary to declare datatype vocabluary to use in @datatype. It is something like @vocab and very similar to RELAX NG datatyping [1] For instance:
>> <div vocab="http://dbpedia.org/property/" datatypeLibrary="http://www.w3.org/2001/XMLSchema#" about="http://dbpedia.org/resource/Albert_Einstein">
>> <span property="dateOfBirth" datatype="date">1879-03-14</span>
>> </div>
>> Instead of:
>> <div prefix="dbp: http://dbpedia.org/property/
>> xsd: http://www.w3.org/2001/XMLSchema#" about="http://dbpedia.org/resource/Albert_Einstein">
>> <span property="dateOfBirth" datatype="xsd:date">1879-03-14</span>
>> </div>
> 
> 
> The 'xsd' prefix is automatically bound to http://www.w3.org/2001/XMLSchema# so there's no need to declare it, which makes the second example:
> 
> <div vocab="http://dbpedia.org/property/"
>     about="http://dbpedia.org/resource/Albert_Einstein">
> <span property="dateOfBirth" datatype="xsd:date">1879-03-14</span>
> </div>
> 
> (Actually if you use <time> then I believe it's
> 
> <div vocab="http://dbpedia.org/property/"
>     about="http://dbpedia.org/resource/Albert_Einstein">
> <time property="dateOfBirth">1879-03-14</span>
> </div>
> 
> because the date will be sniffed as being a date, but that's beside your point.)
> 
> Given, as you say, that most of the time datatypes are XML Schema datatypes, I think it's simpler to just rely on this built-in prefix rather than introducing a @datatypeLibrary attribute. This example shows that it can be shorter, and built-in prefixes have the advantage of never getting left behind during copy-and-paste.
> 
> My 2p.
> 
> Jeni
> -- 
> Jeni Tennison
> http://www.jenitennison.com
> 
> 


----
Ivan Herman, W3C Semantic Web Activity Lead
Home: http://www.w3.org/People/Ivan/
mobile: +31-641044153
FOAF: http://www.ivan-herman.net/foaf.rdf

Received on Wednesday, 23 November 2011 09:12:53 UTC