[tdwg-tag] class design, generalization, L(O)D

Peter Ansell ansell.peter at gmail.com
Tue Nov 16 05:18:20 CET 2010


On 16 November 2010 13:45, Aaron Steele <eightysteele at gmail.com> wrote:
> Hi Peter,
>
>> The data that is generated using microdata looks to be very similar to
>> RDFa, except that the properties are not qualified as URIs, and they
>> seem to be limited to the properties available in the same location as
>> the type definition due to the lack of qualification.
>
> Interesting. So are there things that RDFa can do that microdata
> cannot do? Can you give an example?
>

I think I missed something when I first read through the Microdata
draft, or it was updated in the last few hours, as it has todays date
on it. On closer reading, there can be arbitrary URL's for properties,
so there is no hassle there actually. The examples I was reading
through before I went to the spec only used simple property names from
the vocabulary matching the type of the item.

There is also a good specification of how to generate RDF from
Microdata which includes some common Dublin Core terms as well known
predicates for different parts [1]. However, you could write an RDFa
parser that added the extra triples without having to create a new
spec called Microdata.

I still don't think I would use Microdata, as it seems to duplicate
the RDFa spec that has already been standardised by the W3C (since
2008!), even though technically it is for XHTML and not technically
HTML (yet). The examples in [2] and [3] seem so similar that it seems
like a waste of energy to recommend two slightly different ways to do
the same thing.

[1] http://dev.w3.org/html5/md/#rdf
[2] http://www.w3.org/TR/xhtml-rdfa-primer/#id84801
[3] http://dev.w3.org/html5/md/#names:-the-itemprop-attribute


More information about the tdwg-tag mailing list