abw at wardley.org
Thu Feb 14 09:19:29 GMT 2008
Jonathan Peterson wrote:
> I've been playing with some RDF stuff, and notice a small clutch of Perl
> modules in that direction. RDF::Query seems like the best, but I wondered
> if anyone else was playing in this area.
Implemented in C with Perl bindings. It can use a bunch of different back end
stores including DB, MySQL, Postgres, etc. It comes with an RDF parsing
toolkit (raptor) and a query engine (rascal) which supports both SPARQL and
> I'm slowly coming round to the idea that RDF might be quite good.
RDF is conceptually simple, elegant and expressive. It reminds me of LISP. You
can create some very interesting data models that blow their noses in the
general direction of "traditional" relational databases.
I once wrote a semantic (with a small 's') knowledge engine using a homegrown
RDF store (before Redland was around). With just a single starting axiom
("is") I was able to bootstrap a conceptual universe full of skateboarding
badgers. I got to the point of being able to describe facts like:
buses are big, red vehicles
vehicles have wheels.
skateboards have wheels
skateboards are sometimes red
badgers like riding skateboards
badgers like small things
badgers dislike big things
badgers like red things
It was only a prototypical proof-of-concept so wasn't of much practical use,
unless you wanted help in choosing a suitable present for a badger (hint: red
skateboards go down well, but a bus is too large to fit in the sett). But it
did illustrate some of the interesting stuff you could do with RDF.
However, I suspect that it'll take another 20 years for it (or something
like it) to become widely used. The traditional schema-based database
paradigm is too firmly entrenched in our collective lack-of-intelligence.
So the Semantic (with a large 'S') Web probably isn't happening any time
soon... but that shouldn't stop you from creating your own semantic
More information about the london.pm