[microformats-discuss] Discovery of microformats

[Danny Ayers]

On 7/13/05, Eran <limbo at speakeasy.net> wrote:
> 
> > I believe Brian Suda was working on a parser/validator for xmdp.
> >
> > However, as I'm sure you've noticed, the real guts of an xmdp
> > are the
> > prose descriptions of the elements. So, just like other
> > microformats,
> > xmdp is for "humans first and machines second."

Well yes, that's a nice slogan, but I've yet to see a microformat used
without machine involvement. XHTML+CSS can already provide
human-readable representations of information, what I believe
microformats bring to the table is a way of making that information
explicit in a machine-readable form.

> This shouldn't stop us from including more machine useable information
> in profiles. I'd love to see additional information that can be used to
> automate parsing or viewing of the microformat in question.

I don't personally believe XMDP was needed as there is an existing
standard for describing such information, RDF Schema, which can be
rendered in a human-friendly form, e.g. see [1]. But I'm not going to
argue the point, now there is XMDP there's no reason not to use RDFS
in conjunction with it to maximally exploit the available information.
The easiest approach is to associate the XMDP documents with XSLT
which can extract their data as RDF/XML.

The question still remains of how machines are going to use the data.
I'd suggest there's already considerable infrastructure in place in
the form of Semantic Web technologies.  GRDDL [2] can already provide
automatic parsing of microformats (assuming the profile is
conveniently identified using a URI). RDF stores can easily contain
and process data found in microformat docs. Tools like PiggyBank offer
useful ways of using/viewing the information.

But whether you want to use RDF etc doesn't really matter. The aspects
of microformats that make them useful compared to arbitrary scraping -
standardised vocabularies (hReview etc) unambiguously defined (the
profile has a URI) - are also the things that make the information
useful on the Semantic Web. Whether it's "humans first" or we use
computers to process the data, it's humans that benefit.

Cheers,
Danny.

[1] http://xml.mfd-consult.dk/ws/2003/01/rdfs/
[2] http://www.w3.org/2004/01/rdxh/spec

-- 

http://dannyayers.com