Re: RDF tools as workhorse

William Loughborough wrote:

> Jeen Broekstra wrote:
> 
>> ...I'd say that if your primary concern is query performance then 
>> perhaps using a triple store is not the way to go...
> 
> 
> Excuse a lurker's naivete, but I've been under the impression that 
> the whole point of this entire RDF/triples exercise is to deal with a
>  MUCH LARGER set of data than is envisioned in this thread, i.e. the 
> entire Semantic Web taken as one big database.
> 
> The proprietary ("enterprise" is the usual buzzword) considerations 
> are of much less concern to humanity since what's needed is some more
>  universal everyone/everything/everywhere/always connected sort of 
> thing.
> 
> Finding out how many tires are in stock in Akron is an extremely 
> small part of the overall point - am I missing something?

No, excellent point. Of course, the idea of the Semantic Web is not that
all data resides on a single machine, but on many machines connected
over the Web, using things like SPARQL and HTTP to communicate. If you
will allow me to do a bit of 'hand waving': I believe that serendipity
will play a large role in getting this 'huge world wide database'.
Pockets of information that were unconnected are 'discovered' and linked
together, in ways quite likely unforeseen by the original authors of the
information. I expect that many of these pockets will be what you refer
to as "enterprise" data: stock data and price information from some
company selling tires may not be interesting in and of itself, but when
it is exposed and systems enable one to compare the data with that of a
competitor, for example, or perhaps to integrate the information in the
scheduling process of a luxury car manifacturer who in turn uses it to
service its customers in an optimal fashion, things get more interesting.

All that being said, I think that what the original question was about
was not so much 'how to enable the vision of the semantic web', but more
  'what can I do with triples, right now', and I tried to give a
practical answer to that. What we see with our products in a small scale
is, to me at least, encouraging in both respects however, since we
see that RDF enables us to very easily integrate different kinds of
information in a single system and offer an integrated navigation and
search interface over this information. I think that nicely illustrates
a big strength of RDF, even if it is in a corporate environment rather
than on Web-scale.

And oh yes, of course: performance and scalability of triple stores is
being worked on quite hard these days, and I think there is still a lot
of opportunity for improving.

Jeen

Received on Wednesday, 14 September 2005 17:03:27 UTC