Raw

A highlight of This Week's Semantic Web was the release of Tabulator as Firefox plug-in. Although I installed it pretty much straight away, I didn't have time for a play. This morning I was playing with some data and wanted a copy of the latest FOAF schema. After navigating to the RDF file I was surprised to see this:

After a few moments wondering what danbri had been smoking, I realised I was now browsing data. The Tabulator also allows editing of online data (through SPARQL update), which is intensely cool, though I didn't have any joy with the demo Wiki.

There's something psychologically much stronger about the immediacy of this view just appearing, without having to go through the online p(r)oxy version. As it happened I wanted to tweak the FOAF schema a little before using it in Protege, get rid of the non-DL bits - the owl:InverseFunctionalPropertys that are also owl:DatatypePropertys. I could find these reasonably quickly with the Tabulator extension, although skipping over to the online version I was able to get the same info using its SPARQL engine.

While I accept Doug Engelbart's line about it being reasonable that powerful tools may take some learning, being simpler doesn't necessarily mean dumbing down. In the case of Tabulator I think there's still quite a gap between its current user-friendliness and how it could be, without compromising on features. 'fraid I don't have any specific suggestions, except for lots of usability testing...

(Reminds me, I wonder how the Hyperscope project's coming on - presumably with a format dedumbeddownifier, Tabulator could browse data for that)

Feature request-wise, it'd be good for the plugin to notice when alternate representations were available, showing a HTML page as default but maybe lighting up a button to say there's browsable data available nearby - either through conneg, link rel="alternate" or GRDDL. Conneg should work on this blog, as it stands you have to append .rdf to any page URI.

Firefox's oh-so-wrong handling of feeds through a preset stylesheet messes things up for browsing RSS 1.0. But that aside, where there is RSS 1.0 data (as in this blog) a couple of case-specific tweaks to Tabulator might be worthwhile - render dated rss:items newest first, and unescape rss:content-encoded. As it stands you can use it as a clunky sort of feed reader, but for outreach purposes it'd be sweet if you could go to Planet RDF's bloggers.rdf and easily see a list of items.

Dan Connolly's been playing with GRDDL in Javascript, so hopefully that won't be long in coming.

Perhaps a little pipeline could be built in to Tidy pages as necessary, apply well-known transformations based on media type and/or hacky content sniffing to support microformat conventions and other common formats.

Having said all that, Tabulator's rendering of arbitrary data is remarkably good already. I had to make another copy of this Pet Profile to get the right mime type, but the result was worth it:

 

 

@en