Read THIS first.
Je adore le Web Semantique. Seriously though, the semantic web is really where we are headed and I've written about it loads on this blog . The huge influx of unstructured information generated in no small part by the mass adoption of web 2.0 technology has made semantic web tagging/formatting/searching/aggregating/parsing a necessity. The only way for machines to do that is with the data-centric descriptive frameworks outlined by Sir Tim and codified as RDF and OWL by the W3C.
The problem I have with Bilton's blog article and the twitter example in particular is that he's not thinking like a machine. Sure his compiled tweets indicate he likes vespas and minis. But to a machine all that his tweets show is a small relationship between the string "vespa" and the user object "Nick Bilton". Before that data can be semantically leveraged by a machine it needs to be codified (enter RDF and microformats and API's like OpenCalais).
Incidentally, Oracle has many/most of the pieces necessary for this. Oracle Spatial holds and queries billions of RDF triples against pre-defined OWLs. The data warehousing ETL and ELT technologies *might* be able to be brought to bear on collections of unstructured information (e.g. full text search indices). Oracle UCM collects and full text indexes tons of unstructured information (e.g. documents, emails) that can (should!) be outputting URI encoded RDFs for storage and querying. Oracle EPM and the Analytics suite from the BEA acquisition can be brought into the mix to provide not just metadata but usage data which would automatically be aggregated into usage topographies and formulate the foundation for an organically emerging OWL which would then guide a sophisticated recommendation engine.
Just a vision right now but one that the technology supports. You can bet we're going there.
UPDATE: more discussion here
Comments (4)
I need real use cases like Bilton provides to wrap my brain around this.
We should do something on Connect/OraTweet to make it more real.
Drop a dime.
Posted by Jake | December 12, 2008 3:19 PM
Posted on December 12, 2008 15:19
Great vision !
I hope we will get there very soon.
But I would add to our checklist something more : people.
We have the technology but does we have the people to build an integrated solution that wraps this all together in a Semantic fashion ?
Semantic is not just about technology, its about modeling of knowledge and complex information.
Are we developing these skills to leverage such great technologies and build some killer applications ?
We need knowledge engineers to make this appen. What do you think ?
Posted by Davide Gallo | December 19, 2008 5:23 AM
Posted on December 19, 2008 05:23
Aye, there's the rub. I'd like to be one of those people. Haven't seen / heard of much interest from others though I would *love* to be proven wrong on this.
Posted by billy cripe | December 23, 2008 12:14 PM
Posted on December 23, 2008 12:14
Mee too, I mean I'd like to be one of those people who will bring such solutions to the business.
Actually I know some people that have such skills and have been doing this job for years with in research institutions (such as Universities). Today is considered pioneering but I believe that as today we have BI experts, EAI experts and functional experts to support each ERP related process, in the near future we will need these kind of Semantic experts.
Now is time to team up and start building solutions to prove we have the right vision.
Or we can simply provide technology and wait that someone other builds solutions with it...but aren't we missing most of the potential value ?
Happy holidays !
Posted by Davide Gallo | December 23, 2008 4:33 PM
Posted on December 23, 2008 16:33