Excerpts from Nina Jeliazkova's message of Fri Oct 29 15:12:39 +0200 2010:
> On 29 October 2010 15:52, Egon Willighagen <egon.willighagen at gmail.com>wrote:
> 
> > On Fri, Oct 29, 2010 at 2:42 PM, Nina Jeliazkova
> > <jeliazkova.nina at gmail.com> wrote:
> > > If there are no R RDF packages, perhaps a C RDF library could be used ?
> >  R
> > > seems to be able to use different languages easily.
> >
> > I you find an answer, to can answer my question about it on BioStar:
> >
> >
> > http://biostar.stackexchange.com/questions/2069/how-do-i-import-rdf-data-into-r
> >
> > :)
> >
> >
> I've found this note interesting ...
> 
> http://www.biostat.harvard.edu/~carey/hbsfin.html
> 
> Nina

Based on my ruby experience I would say that a pure RDF library
(Rredland does  not look too bad) does not help very much for pure
format conversions from RDF to data frames (unless it supports OWL).
Parsing/serialising triples directly is much faster and uses almost the
same amount of code. The ugly part its to get the OWL-DL right (I would
hate to repeat that exercise in R) - and here pure RDF libraries do not
help in any way. The only advantage of RDF libraries is for querying
data, but it comes at the performance cost of building and maintaining
the RDF graph. And querying can be done either at OpenTox services or
within R's representation.

So for short term success it might be easier to use one of the existing
OpenTox libraries in Java (or Ruby), although such solutions tend to be a
kludge and might be hard to maintain in the long term.

A motivated maintainer could build a much cleaner solution with
integrated conversion between R and OpenTox objects and a proper
integration of curl C libraries (if there are no other R libraries for
that purpose).

Best regards,
Christoph