- From: Dudley Mills <dudley.mills@bigpond.com>
- Date: Sun, 20 Jan 2008 16:50:36 +1100
- To: <semantic-web@w3.org>
- Message-ID: <000001c85b28$64ed4570$0100000a@Varanusprisca>
Grazie Ignazio, > Any (semi)automatic tool for annotation does the same, some of them > actually produce RDF data, some others predate RDF and store the > semantically rich stuff in other formats. One buzzword you can use to > find many of these is Named Entity Recognition. Google for it, you'll > get tons of results. Patent http://www.freepatentsonline.com/7289956.html, one of many, highlights the problem with Named Entity Recognition: our computers still can not process natural language with sufficient precision. We still need to feed them data using narrowly defined languages like XML on which the language of the Semantic Web is built. This may remain the case for decades because context is very important in natural language processing and context is generally not completely obvious when scanning a web page. > A patent over some very simple and/or very old idea is not a > useful service. The best patents are over the simplest most commonly and most frequently used inventions. A common example being spreadable refrigerated butter. A patent over an old (not novel) idea is not worth anything.
Received on Sunday, 20 January 2008 05:52:05 UTC