68

Idea:
Create a soprano backend for the live, online DBpedia RDF store. From the DBpedia site:

The DBpedia knowledge base currently describes more than 2.6 million things, including at least 213,000 persons, 328,000 places, 57,000 music albums, 36,000 films, 20,000 companies. The knowledge base consists of 274 million pieces of information (RDF triples). It features labels and short abstracts for these things in 30 different languages; 609,000 links to images and 3,150,000 links to external web pages; 4,878,100 external links into other RDF datasets, 415,000 Wikipedia categories, and 75,000 YAGO categories.

Benefits:
Applications would have quick, api-level, semantic access to wikipedia information. Examples include:
- A runner that would give perfects matched to queries like: "French films" or "Senators from California" or "German musicians born in Berlin" or "What is a quark".
- Multimedia apps could provide info on music, film, artist, actors, producers directors, other film by same director...
- Semantic text scanning would recognize wide variety companies, places, people, etc.
- Location-aware apps could provide in depth information on your current city, country or region.

Note that this new backend would provide a tool for the developer. In order for users to see benefits, developers would need to take advantage in their apps. Between Nepomuk and this backend developers would have access to both local and global knowledge bases in via the same powerful semantic layer.

(P.S. Call me crazy, but I can just imagine a speech recognition engine that generates text for queries. This is simply match a text string... It\'s our desktop "knowing" what your asking for because is it "understands", for example, that Sony is a Corporation with a CEO who is a person with a birth date and a place of birth that is a city or country that is a parliamentary democracy with a head of state who is person with a name and has a birth date... Crazy...)

jamboarder wrote:Create a soprano backend for the live, online DBpedia RDF store.

DBpedia is great, though I am not sure if it is their datastore that should be accessed. DBpedia is more a semantic web experiment and demonstration. Rather Soprano should have knowledge about and access to all existing RDF providers and do itself what DBpedia does. Of course, right now DBpedia has to extract and salvage many of its data (as from Wikipedia for example), but some already provide real RDF data (like Musicbrainz). In the future all data will be offered as RDF triplets and machines will be able to autonomously communicate which each other and provide the user with the information s/he wants.

jamboarder wrote:Call me crazy, but I can just imagine a speech recognition engine that generates text for queries. This is simply match a text string... It's our desktop "knowing" what your asking for because is it "understands", for example, that Sony is a Corporation with a CEO who is a person with a birth date and a place of birth that is a city or country that is a parliamentary democracy with a head of state who is person with a name and has a birth date... Crazy...

This is not crazy at all. Actually this is exactly what the semantic web/desktop is all about. It's the next step towards artificial intelligence, and, once the semantic web has sufficiently evolved, your example is just one possiblity of many.

tremendous great idea!

I wrote:[...] but some already provide real RDF data (like Musicbrainz). In the future all data will be offered as RDF triplets [...]

Guess I have to correct myself as I've seen that Musicbrainz doesn't serve RDF anymore, they say it's depreciated.

This is something I'd prefer to be done by Soprano/Nepomuk itself... it would get the data from all these different websites by itself and easily allow developers to add new websites to the capabilities, rather then relying on another independent project. This way, you can guarantee that the information will be freely available and have other services installable through GHNS, if they can't be included for some reason (such as proprietary mapping information or other such cases).

Having worked with DBPedia data for Semantic Web research projects, I would express a few concerns against using it as you suggest, even though the idea is interesting.

A first obstacle would be the query speed of DBPedia itself through the network (a server shall be able to address all requests from KDE users ), and the size of the data would make it hard to allow it to be locally duplicated (which by the way would probably be against the spirit of the Semantic Web). However, the main concern would be the quality of the data, since DBPedia would require a significant cleaning to be both usable and up-to-date.

In my opinion, a full set of different and specialized providers offering an access to high-quality RDF data following a common ontology (or ontologies that can be aligned with each others) would be far better (e.g. a music provider would be able to propose up-to-date data about music tracks, while other providers can concentrate on their respective fields).

Like it was said, DBPedia is more of an experiment to show what would be a semantic data collection, but the Semantic Web is more about "small" interconnected and interconnected data repositories. I reckon that it is more the trend that will be followed in the future by Nepomuk and others...