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

Leigh Dodds ldodds at ingenta.com
Wed Aug 27 09:09:24 UTC 2003


> >  * "classic" FOAF files
> 
> As is probably clear, this is the one I'm interested in. What I have in 
> mind is a set of conventions. And maybe that's all it is. eg

This isn't a million miles away from what the current FOAF-a-Matic 
generates, so the profile is already in wide-spread use. My latest 
code does have more permissible variations, in-line with building 
it around Jena.

> - Only one main Person with Person as a top level node

This works for the FOAF-a-Matic because its about 1 person authoring 
a FOAF file. Same applies to any "publish my details" application.

>    - This Person is assumed to be the author/owner of the data unless
>      otherwise stated.

Could we not use foaf:maker here? If not then you'll have to have some 
other means of stating that the person, or someone else is the author/owner 
of the data. As foaf:maker goes in a separate rdf:Description element this 
means the file isn't entirely hierarchical, but this is only a single cross-reference 
and no worse than dealing with ID/IDREF in plain ole XML. That rdf:Description 
element gives you a place to hang things like admin:generatorAgent as well.

You also suggest avoiding RDF/XML shortcuts, can I propose that you do 
allow for:

<foaf:Person foaf:mbox_sha1sum="..."/>

as well as

<foaf:Person>
  <foaf:mbox_sha1sum>...</>
</foaf:Person>

It's no great pain to support this kind of short-cutting of non-repeated 
properties using plain ole XML tools. And as some RDF libraries (e.g. Jena) 
don't give the programmer control over which usage is generated (AFAIK, anyway) 
it would slightly broaden the range of tools and data to play with.

Cheerio,

L.



More information about the foaf-dev mailing list