- From: Dan Brickley <danbri@w3.org>
- Date: Tue, 25 Jul 2000 16:16:18 -0400 (EDT)
- To: www-rdf-interest@w3.org
(what Guha _meant_ to say was that this is a semistructured database for XML-powered Web apps that use an edge-labelled graph datamodel. Like RDF :-) <ducks/> However described, it's a nice bit of work and (judging by the linux binary) a good step towards shrinkwrappability of RDF database/query/parser tools. So please send Guha lots of bug reports (and fixes...). The query language aspect is interesting, and perhaps worth discussing here. The current system uses SQL-ish textual query strings rather than the use-markup-for-all-structure philosophy. Other systems I've seen have a similar but not the same approach. It'd be good to have some basis for interchanging queries (and inference rules) between applications like RDFdb. Which raises a bunch of issues, one of which i'm currently puzzling over: how do we represent query variables to RDF? This cropped up in the DanC/Jos discussion of inference rules recently[1], where Dan quite reasonably objected to "recognizing variables by their spelling". I really don't know which approach makes sense for this right now, but still I find 'var:' appealing, much as 'phone:', or 'java:' etc can be used for homegenous regions of the URI namespace. That said, claiming a toplevel URI scheme is serious business; maybe just using http://example.com/variables/foo would be just as good. DanC's model is intriguing, but doesn't really address the issue of which URIs do we use to identify the variables in some representation of a query. (I think his answer is 'any will do'...) Has anyone else come up with models for representing variables over RDF? Might be fun to write translations to/from Guha's RDFdb language, for example... --dan [1] http://lists.w3.org/Archives/Public/www-rdf-interest/2000Jul/0028.html On Tue, 25 Jul 2000, Guha wrote: > Hi, > > I'd like to announce the first release of RDFdb > (http://www.guha.com/rdfdb/). > > RDFdb is an open source (under the MPL) scalable triple database server > with a > SQL-like front end. Its written in C with a perl client. > > You can download a linux binary or the source from the above url. > > I'd appreciate any comments, feedback, bug reports, bug fixes, ... > > thanks, > > Guha >
Received on Tuesday, 25 July 2000 16:16:22 UTC