Re: Test of Independent Invention: RDF

On 05/11/2015 03:47 PM, Harry Halpin wrote:
> 1) The largest deployment barrier is probably IMHO the lack of a fast,
> scalable, open-source triple-store that can replace Postgre or the
> noSQL solutions like MongoDB people know and use.

Interesting.  I have always viewed performance as a question of learning 
how to properly use the technology -- not so much a fault of the 
technology or implementation itself.  Just as one learns not to join 
across 13 tables in a relational database, one learns not to issue 
SPARQL queries that would similarly cause combinatorial blow-up.  There 
is a learning curve with any technology, and the development community 
as a whole is much farther along that learning curve with relational 
than semantic technology.

Still, it would be interesting to see direct performance comparisons 
between relational queries and SPARQL for the same kind of task, 
competently designed in both relational and RDF approaches, to quantify 
how much the RDF world is lagging in performance.

I'm also curious to know what others think are the largest deployment 
barriers for RDF.  Why haven't people used it when you thought they 
should?  Or why have they abandoned RDF after trying?

David Booth

Received on Monday, 11 May 2015 20:29:17 UTC