- From: Tim Berners-Lee <timbl@w3.org>
- Date: Tue, 5 Jun 2001 21:45:59 -0400
- To: "Jonathan Borden" <jborden@mediaone.net>, "Peter Crowther" <peter.crowther@networkinference.com>
- Cc: <www-rdf-logic@w3.org>
Jonathan Borden wrote, "I strongly suspect that a relational db implemention based on n-tuples will very much outperform an implementation using a transformation to triples" I don't think one can generalize this. Different applications will have data with very different properties. One may have very dense 5-ary data which will most efficiently be stored in a RDB table with 5 columns, but another, maybe using the same Properties, might be so horrendously sparse that representation as triples would fact saves memory on your PDA. The sorts of queries you want to do on the data affect the optimum storage too, of course. A pint of Sam Adam's Ale to the first person to implement an RDF store which automatically optimises the underlying database representation as a function of the statistical properties of the requests made on it and/or the data stored in. Tim ----- Original Message ----- From: "Jonathan Borden" <jborden@mediaone.net> To: "Peter Crowther" <peter.crowther@networkinference.com> Cc: <www-rdf-logic@w3.org> Sent: Tuesday, June 05, 2001 7:52 AM Subject: Re: Cheap memory [was RE: rdf as a base for other languages] [...]
Received on Tuesday, 5 June 2001 21:46:11 UTC