I just discovered the most amazing thing. It's a jabber/xmpp sparql interface, called xOperator.
It sits between you, the end user, and multiple other agents and datastores. It uses XMPP (Google talk!) to interact with you.
You say things like:
me: add ds DBPedia http://dbpedia.org/sparql
me: query SELECT ?name WHERE {
?musician skos:subject <http://dbpedia.org/resource/Category:Australian_musicians> .
?musician foaf:name ?name .
}
bot: (Alright, this is dummied up, but I *did* have it working with other stuff before)
"Ben Lee"@de
"Billy Thorpe"@de
"Brody Dalle"@de
"Charlie McMahon"@de
"Craig Nicholls"@de
"Daevid Allen"@de
"Delta Goodrem"@de
"Fritz Hart"@de
"James George Thirlwell"@de
"Joe Dolce"@de
"Johnny Diesel"@de
"Kasey Chambers"@de
"Kevin Johnson"@de
"Natalie Gauci"@de
"Percy Aldridge Grainger"@de
"Peter Garrett"@de
"Peter Allen"@de
"Rick Springfield"@de
"Richard Franklin"@de
"Sophie Monk"@de
"Steven Kilbey"@de
"Stevie Wright"@de
"Tina Arena"@de
"Toni Collette"@de
me: add template "Tell me about *" "PREFIX owl: <http: org="" 2002="" 07="">
PREFIX xsd: <http: org="" 2001="">
PREFIX rdfs: <http: org="" 2000="" 01="">
PREFIX rdf: <http: org="" 1999="" 02="">
PREFIX foaf: <http: com="" foaf="" 1="">
PREFIX dc: <http: org="" dc="" elements="" 1="">
PREFIX : <http: org="" resource="">
PREFIX dbpedia2: <http: org="" property="">
PREFIX dbpedia: <http: org="">
PREFIX skos: <http: org="" 2004="" 02="" skos="">
SELECT ?name WHERE {
?thing skos:subject <http: org="" resource=""> .
?thing foaf:name ?name .
}"
bot: template added
me: Tell me about German_musicians
me: Tell me about German_artists
bot: Store dbpedia answered:
name
"A. R. Penck"@de
"Adolf Ziegler"@de
"Adolph Menzel"@de
"Albrecht Altdorfer"@de
"Alfred Harth"@de
"Alfred Rethel"@de
"Alfred Mahlau"@de
"André Butzer""@de
"Bruno Paul"@de
"Bärbel Bohley""@de
"Carl Grossberg"@de
"Carsten Höller""@de
"Charles Crodel"@de
"Christian Möller""@de
"Christoph Meckel"@de
"Christoph Ruckhäberle""@de
"Cosmas Damian Asam"@de
"Dieter Grossmann"@de
"Dieter Roth"@de
"Eberhard Bosslet"@de
"Eberhard Havekost"@de
"Egid Quirin Asam"@de
"Elsa von Freytag-Loringhoven"@de
"Ernst Deger"@de
"Hugo Höppener""@de
"Franz Joseph Spiegler"@de
"Franz Wilhelm Seiwert"@de
"Franz Xaver Feuchtmayer der Jüngere""@de
"Franz Xaver Feuchtmayer"@de
"Frieder Nake"@de
"Fritz Maurischat"@de
"Gerd Aretz"@de
"Gerd Arntz"@de
"Gert Heinrich Wollheim"@de
"Gert Neuhaus"@de
"Hans Haacke"@de
"Hannah Höch""@de
"Hans Grundig"@de
"Hans Richter"@de
"Heiko Daxl"@de
"Heinrich Aldegrever"@de
"Herma Auguste Wittstock"@de
"Herbert Holzing"@de
"Horst Janssen"@de
"Horst Antes"@de
"Ignaz Günther""@de
"Jankel Adler"@de
"Hans Arp"@de
"Jim Avignon"@de
"Johann Peter Melchior"@de
"Johann Baptist Straub"@de
"Johann Joachim Kändler""@de
"Johann Joseph Christian"@de
"Johann Melchior Dinglinger"@de
"Johann Baptist Zimmermann"@de
"Johann Michael d.J. Feuchtmayer"@de
"Johannes Baader"@de
"Jonas Burgert"@de
"Joseph Anton Feuchtmayer"@de
"Gotthard Johnny Friedlaender"@de
"Jonathan Meese"@de
"Joseph Beuys"@de
"Karin Ulrike Soika"@de
"Karl von Fischer"@de
"Karl Hubbuch"@de
"Klaus Herlitz"@de
"Konrad Peter Cornelius Klapheck"@de
"Kurt Wegner"@de
"Leo Kahn"@de
"Franz Karl Leopold von Klenze"@de
"Lore Lorentz"@de
"Ludwig Fahrenkrog"@de
"Martin Kippenberger"@de
"Manfred Mohr"@de
"Manfred W. Jürgens""@de
"Maria Innocentia Hummel"@de
"Mariele Neudecker"@de
"Werner Mathias Goeritz Brunner"@de
"Mary Bauermeister"@de
"Max Ernst"@de
"Michael Buthe"@de
"Matthäus Günther""@de
"Michael Sowa"@de
"Otto Eckmann"@de
"Otto Hupp"@de
"Ottmar Hörl""@de
"Otto Pankok"@de
"Otto Piene"@de
"Peter Vischer"@de
"Peter Dreher"@de
"Rebecca Horn"@de
"Ruben Talberg"@de
"Rosemarie Trockel"@de
"Sandra Rauch"@de
"Silvia Quandt"@de
"Thomas Eller"@de
"Thomas Schütte""@de
"Tilman Riemenschneider"@de
"Tomma Abts"@de
"Torsten Amft"@de
"Uwe Laysiepen"@de
"Veit Stoß""@de
"Virgil Solis"@de
"Wilfried Behre"@de
"Wilhelm Heine"@de
"Wilm Weppelmann"@de
"Wolfgang Joop"@de
"Wolfgang Müller""@de
"Wolfgang Tillmans"@de
"Evi Moechel"@de
(10:22:30 PM) clockwerx@livejournal.com/Home: Tell me about Australian_musicians
(10:22:32 PM) daniel.oconnor: Store dbpedia answered:
name
"Ben Lee"@de
"Billy Thorpe"@de
"Brody Dalle"@de
"Charlie McMahon"@de
"Craig Nicholls"@de
"Daevid Allen"@de
"Delta Goodrem"@de
"Fritz Hart"@de
"James George Thirlwell"@de
"Joe Dolce"@de
"Johnny Diesel"@de
"Kasey Chambers"@de
"Kevin Johnson"@de
"Natalie Gauci"@de
"Percy Aldridge Grainger"@de
"Peter Garrett"@de
"Peter Allen"@de
"Rick Springfield"@de
"Richard Franklin"@de
"Sophie Monk"@de
"Steven Kilbey"@de
"Stevie Wright"@de
"Tina Arena"@de
"Toni Collette"@de
Big whoop you say.
Well, I've been working on both an XML_GRDDL package (php5), and the relevant GRDDL transformations to turn
LIXI Valuation documents into the relevant RDF/XML representation.
I'm also re-deploying work's internal Jabber server in the not too distant future.
So:
1. We handle a pretty big chunk of all residential valuations within Australia.
2. A lot of that gets described adequately in LIXI
3. There's now or soon to be a GRDDL transformation or two for that
4. And a GRDDL parser for PHP
5. Add in xOperator
6. Add in robust jabber services
7. Mix in a few microformats as well
Viola:
My workplace is half a step away from becoming a
knowledge enterprise as well as a valuation panel manager.
That's not bad for a few days off!
Imagine these future scenarios:
Working smarter with internal data and external clients:Internal xOperator to answer questions like "show me similar valuations we've done to address xyz", and "show me everything valuer X has completed over $1million"
Paid commercial access to a SPARQL service as well, for larger clients who want to better understand the marketplace
Paid commercial access to an xOperator style service for smaller clients, who just want a couple of pre-built, on demand answers
Publishing old (> 12-24 months) LIXI data on the web
Merge our data up with other countries, like the US, for trend analysis
Provide anonymized public data, with http basic auth. to protect names, addresses, privacy
Merge diverse sets of data - ie, imagine EuroStat data for Australia, but with an atomic level of detail available about the raw input
Provide semi-automated valuations based on known data
Suburb X has a population of Y and an area of Z (db/wikipedia), factor that into calculations
