W3C home > Mailing lists > Public > www-rdf-interest@w3.org > August 2000

Re: Announcing RDFdb

From: Dan Brickley <danbri@w3.org>
Date: Fri, 18 Aug 2000 14:07:41 -0400 (EDT)
To: www-rdf-interest@w3.org
Message-ID: <Pine.LNX.4.21.0008181357250.5781-100000@tux.w3.org>

On Fri, 18 Aug 2000, Guha wrote:
>  I have plans of supporting reification. Just haven't done it yet.
> 
>  The indexes are currently built on b-trees provided by Sleepycat.
>  I am working on a "nested hash" structure which is more appropriate
> for graphs. The "join algorithm" is fairly straightforward and
> only uses conjunct ordering for optimization. I plan to implement
> a lot of the optimizations used in cyc, lore, etc.

A bit of an aside...

If you're swapping out the Sleepycat/BerkeleyDB layer, maybe this mightn't
be of such interest... Dave Beckett pointed me yesterday to this
"MaxSQL" article on MySQL.com:
	http://www.mysql.com/news/article-28.html

	MaxSQL is a MySQL distribution compiled with Sleepycat software's
	BerkeleyDB support for transactions.
                       MaxSQL will be released under the GNU GPL licence
	as soon as BerkelyDB transactional tables are sufficiently
                       stable and functional to pass our test suite. The
	release is scheduled to the end of August. Users can evaluate
                       the current state of our transactional code by
	installing the lastest 3.1 version of BerkeleyDB available from
                       Sleepycat, and then compiling MySQL on their
	systems with a special option (--with-berkeley-db). More
                       information on MaxSQL will be released on this site
		later. 


Since a bunch of us have been looking at ways of stuffing RDF into
either/both BerkeleyDB and the main opensource RDBMS systems (PostgreSQL,
MySQL) I thought this worth a mention. I know people have grumbled about
the b-tree stuff in MySQL (not to mention lack of transactions), perhaps
this will offer some improvements?

That said I'm inclined to agree with Guha here that ultimately we need
something that's down-to-the-metal optimised for edge-labelled
squiggly graphs. In the meantime, we can get on with discussing APIs,
query languages etc and have perfectly reasonable initial implementations
built on top of SQL and BDB...

--danbri
Received on Friday, 18 August 2000 14:07:41 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:51:44 GMT