Excerpts from Nina Jeliazkova's message of Mon Oct 05 11:28:25 +0200 2009:
> I would suggest PMML as export/import format where possible, in order to
> allow for interoperability with different software solutions. We dont'
> want OpenTox platform to provide only  closed custom solutions, don't we?
> 
> Ideally would be good if the internal XML formats for models and
> everything else follows some established formats, e.g. Resource
> Description Framework for example.  I realize it might be too demanding
> to move to RDF for API 1.1 - we need to keep the deadlines we have set -
> but let's think about it for the next API 1.2.

I had a quick look at RDF and from what I understand up to now it seems
to be well suited for our purpose - here are some use cases from the
intro at http://rdfabout.com/quickintro.xpd:

    * You want to integrate data from different sources without custom
      programming.
    * You want to offer your data for re-use by other parties
    * You want to decentralize data in a way that no single party "owns"
      all the data.
    * You want to do something fancy with large amounts of data (browse,
      query, match, input, extract, ...), so you develop (or re-use) a
      generic tool that allows you to do this on top of the RDF data
      model (which has the advantage of not being tied to a proprietary
      data storage/representation technology, like a database dialect).

Seems to be exactly what we need.

I like that RDF is a data model, not a format. I dislike the clumsy XML
representation (but there is also N3 and its derivatives) and it seems
that RDF Ruby libraries are less than mature.

Best regards,
Christoph