see also: JS-RDF Testbed installation, full XML demo (in progress), samples
Right now, what you see is what you get. It's fun, simple, mostly works. Jan might be persuaded to add more features to the underlying engine, or he might decide the whole thing is silly... To use, best bet is to take a look at an existing installation and copy the prolog.js and rdfquery.js files locally.
2000-07-21: Jan has added some new facilities to the prolog version (cut, builtins etc). I need to merge the RDF addons back in...
[Prolog] Love it or hate it, you just can't program declaritavely in it. This subset does support declarative programming by dint of removing (read: not implementing) every non-logical construct (and a few others besides) from Prolog.
The most recent testbed is the ILRT js-rdf page.
This includes both Jan's prolog, DanB's little RDFish wrapper for it, and the MinXML parser. The RDF/prolog user interface appears at foot of the page. The demo can eat RDF in a homegrown MinXML serialisation syntax (simpler than the standard RDF syntax) as well as the pseudo-prolog RDF format hacked into Java's code).
There's also a version installed on the W3C site as part of a piece on web history (rdfs tests, longer writeup).
The ATAGDemo version is different again (data only -- all the W3C-hosted copies use same .js files), and shows the use of RDF query and simple rules to reason about the capabilities of WWW authoring tools. Hacked together in collaboration with Charles McCathieNevile of the W3C's Web Accessibility Initiative (WAI); idea being that we have a distributed database in RDF that describes WWW software tools in terms of the WAI content authoring guidelines.
The only other version to date (apart from Jan's 5K submission of the (compressed) prolog engine itself) is the little paper on the DESIRE site, Metadata, Web caching and URI resolution. The paper includes an embedded RDF query demo showing how we can use metadata to reason about URIs and the location(s) of the resources they describe.
Dan Connolly has been experimenting with XSLT for semantic web screenscraping (see scribe demo for references), which provides RDF data that can be queried and processed in js-rdf.