I hardly got any response to the launch of my JAM*VAT structured aggregator tool. Either that means that nobody's got a use for it, or they just don't understand what does. I'm hoping it's the latter, so I thought I'd post some things to try with the demo installation. Here's a first stab:

Getting a list of titles, in order, from an RSS1.0 feed.

This example illustrates using JAM*VAT as an RDF store. The really cool thing is that jamvat maintains the order of statements in a graph (something that native RDF stores don't tend to do as RDF is unordered) which compliments RSS1.0's implicit item ordering.

Here's the steps:

  • Import Edd Dumbill's RSS1.0 feed using the import page with a source URL: http://usefulinc.com/edd/blog/rss and format RDF. (actually, check the 'graphs' page first - somebody may have already imported it)
  • Check that the feed looks ok - go to text search and search for e.g. 'edd dumbill' or maybe 'edd dumbill channel', then follow a link and browse around the data a bit
  • Now execute the following structured query using the structured query page:
    select ?title
    where (?channel title ?chantitle)
          (?chantitle _SUBSTRS_ "Edd Dumbill")
          (?channel items ?seq)
          (?seq ?node ?item)
          (?item title ?title)
    This query basically looks for a channel with a title containing the words 'Edd' and 'Dumbill', and then does sparql-style triple-matching to deliver the item titles. Cool eh!

N.B. Every so often the textdrive server running the software demo seems to pause. I know it's the server because I can't ssh to it when this happens. It appears to recover after a couple of minutes though so if this happens please be patient.