Re: Hello, n3

Hi all,

Thanks for initiating this, Nathan.

Myself, I’ve been working with RDF off and on since about 2009 or 2010 when
I was part of a project about data catalogues and another about library data.

For the past several years I’ve been part of the Synthsys Centre for Synthetic
Mammalian Biology. One thread of research  involves simulation of molecular
interaction networks. This is done by describing the molecular interactions in a
specialised programming language. These programs can be (A) annotated with
RDF to indicate the biological entities involved or (B) entirely generated from a
high-level RDF (Turtle) description of a genetic circuit.

The latter exercise exposes a usage pattern of inference rules that I would like
to highlight. A real person writes the input, and for specialist tasks like this,
writing the input in Turtle masquerading as a domain-specific language works
pretty well. But we don’t want to over-burden the user, so it is better to have 
them write as little as possible and to provide rules for deriving all of the extra
facts that are needed to actually produce a running program. This is very
different from the practice in data publishing which is to assume nothing about
the ability of data consumers to be able to do inference, so it is best to include
all kinds of redundant triples just in case. Rather economy of representation is
an explicit goal and using rules, usually written in N3 and often ad-hoc, is 
very effective for producing whatever else is needed.

Most often when working with RDF I use the Python rdflib and FuXi which
can run N3 rules and from time to time I have made some very minor
contributions to these.

Cheers,

William Waites | wwaites@inf.ed.ac.uk
Laboratory for Foundations of Computer Science
School of Informatics, University of Edinburgh