My.Opera SPARQLs
Tuesday, 20. December 2005, 16:18:22
In recent discussions on how the Semantic Web might happen people have suggested that exposing a lot of data was one of the things that is necessary. I have for a long time claimed that this in turn is held up by the lack of a common trust/security infrastructure, since a lot of data is somewhat private and people aren't prepared to simply turn it over to the universe.
On the other hand, there is a lot of public data. Kjetil Kjernsmo has been working on My.Opera, and thanks to him Opera recently launched the My.Opera SPARQL interface which I believe is the first public SPARQL interface to a commercially-backed datastore - the information that people publish on my.opera. With a quarter of a million (or so) registered users, whose data ranges from only a made-up name to a substantial profile, this is a pretty big set. At the moment the data gets updated weekly, although hopefully that will change as the service is developed.
So what is this actually good for? Well, you can link information to other RDF information. You can write a query to find out who writes about your favourite topic and claims to be based in Oslo (or Atlantis...). You can link your travel itinerary to people whose interest includes "meeting RDF geeks" and who live somewhere you have to stop over for the day. You can find out whether someone who speaks your language has written about widgets.
It also means that the people who put the data there in the first place can get it and do whatever they want with it, rather than only allowing them to use it locked up inside our software - something that most other community sites don't really get yet.
(And it got Arve blogging on RDF, which is a nice thing to see
Maybe he'll get motivated to learn even more about it...)
Like the Web in general, the exciting thing about the Semantic Web is not when there are a few collections of data, but when there are a few collections of stuff that you can link together. Earlier hypertext systems generally didn't provide the glue for linking to different stuff that made the Web, once the critical mass appeared, such an amazing resource (and drove the development of search engines - something that SPARQL is a key part of for the Semantic Web).
It's an experimental service that we launched. We hope it will help the development of the Web in general, providing real test cases to anyone that wants to develop something cool. It's nice to do something first, especially when we are not (yet, anyway ;-) ) one of the more famous Semantic Web development groups. If it takes off, there is more we could do to make it easy for people to develop new applications - stuff for Opera platform or wanting data to feed to existing applications like htmlnaut, or whatever the collective minds of humanity can come up with. And it's fun stuff to play around with...
On the other hand, there is a lot of public data. Kjetil Kjernsmo has been working on My.Opera, and thanks to him Opera recently launched the My.Opera SPARQL interface which I believe is the first public SPARQL interface to a commercially-backed datastore - the information that people publish on my.opera. With a quarter of a million (or so) registered users, whose data ranges from only a made-up name to a substantial profile, this is a pretty big set. At the moment the data gets updated weekly, although hopefully that will change as the service is developed.
So what is this actually good for? Well, you can link information to other RDF information. You can write a query to find out who writes about your favourite topic and claims to be based in Oslo (or Atlantis...). You can link your travel itinerary to people whose interest includes "meeting RDF geeks" and who live somewhere you have to stop over for the day. You can find out whether someone who speaks your language has written about widgets.
It also means that the people who put the data there in the first place can get it and do whatever they want with it, rather than only allowing them to use it locked up inside our software - something that most other community sites don't really get yet.
(And it got Arve blogging on RDF, which is a nice thing to see
Maybe he'll get motivated to learn even more about it...)Like the Web in general, the exciting thing about the Semantic Web is not when there are a few collections of data, but when there are a few collections of stuff that you can link together. Earlier hypertext systems generally didn't provide the glue for linking to different stuff that made the Web, once the critical mass appeared, such an amazing resource (and drove the development of search engines - something that SPARQL is a key part of for the Semantic Web).
It's an experimental service that we launched. We hope it will help the development of the Web in general, providing real test cases to anyone that wants to develop something cool. It's nice to do something first, especially when we are not (yet, anyway ;-) ) one of the more famous Semantic Web development groups. If it takes off, there is more we could do to make it easy for people to develop new applications - stuff for Opera platform or wanting data to feed to existing applications like htmlnaut, or whatever the collective minds of humanity can come up with. And it's fun stuff to play around with...
WIDGETIZE