-=( In Between )=-

While converting the metadata in our repositories to “RDF”, we (my student Florian Kunneman and I) wanted to express a few simple relations between authors. Initially we thought we could use foaf:knows for people who co-authored a document. But that’s plain silly, of course. We can use foaf:knows, but would need to subclass it.  We would need  something like the extension made by Eric Vitiello’s extension to FOAF to express family relations (you know: friendOf, parentOf, siblingOf, etc).

We would like to extract relations of the following sort:

Author1 cites Author 2
Author1 mentions Author2
Author1 hasCoAuthor Author2
etc..

A problem might be the fact that such relations have to be derived from publications. Take the first example. It would in fact be a derivation from other relations:

Author1 isAuthorOf Document1
Document1 hasCitation Document2
Author2 isAuthorOf Document2
it then could follow that
Author1 cites Author 2

I wonder how I can find work relevant to this question. Perhaps we first need to define relations that can exist between persons and documents, and from that deduce relations between persons (the same goes for documents of course). What are the relations between authors that could interest us?  It may sound silly, perhaps it is, but I don’t know how to proceed (well we could make our own AOAA -Author Of A Author- vocab, but that again sounds silly).

I played around with Google Wave a bit. I didn’t get it first, but after a while I saw its charms, well, potential charms. We finally seem to have a non clumsy tool to collaborate and communicate, the latter perhaps more than the former.

It may help to streamline thoughts on how to proceed if we want our (digital) libraries to become part of the semantic web. I started a wave therefore, and posted the following message.

——

Working from the assumption that the library should become part of the semantic web, what should we do? Is the following a reasonable set of things to do?

1) Identify what we have, and give it a URI (meaning that authors, documents (objects), keywords, and concepts should get a URI

2) deposit the URI’s in a registry, with a minimal set of triples (relate URI of documents to URI of authors, etc…): what is the minimal set?

1 and 2 are data level requirements, the following are service level

3) use OAI-ORE to define compound objects

4) define owl:sameAs relations between URI’s when needed (what is the workflow here?)

5) Which vocabularies/ontologies can/should be used for which services. Do we need a service catalogue first?

what else?

——

Now, if you are interested in active participation, let me know