[rdfweb-dev] Re: Syntactic profiling (FOAF document formats)

Danny Ayers danny666 at virgilio.it
Thu Aug 28 13:07:55 UTC 2003


> * you only get reusability benefits because the format is forked in the
> first place
>
> Using an RDF parser to process FOAF is *easier* then regex land.
> Especially if you go further than just using a parser and use a toolkit
> like Redland, Drive or Jena because then your data model is done for you
> as well.
>
> Really, let's not make this harder than it should be.

For me, these are the key points. I have periodically drifted into thinking
that a plain XML profile of FOAF might be a good thing because it might
encourage more people to produce the data. But it is deceptive - once you
accept the RDF model (i.e. use a toolkit) then RDF/XML is easier to use than
XML. An XML vocabulary in isolation is simple to handle, but there are major
problems when it comes to mixing vocabularies (basically, you can't).

Having to deal with RDF/XML *plus* a pure-XML syntax is harder than either
of these - and that's the uncomfortable situation with RSS. I really think
we should avoid going there with FOAF.

As an aside, it's interesting how far plain-XML RSS has got without any
standard XML Schema definition (or DTD for that matter). I think this
suggests that FOAF XML schemas would bring little gain in practice. The
syntax checking of RDF parsers seems to be perfectly adequate for spotting
dirty data. I'm not sure about datatype checks, but so far I've not
encountered any evidence that they are likely to cause problems.

Cheers,
Danny.




More information about the foaf-dev mailing list