Ahem. How is this news? In fact, we (AI students) had to write a NLP (Natural Language Parser) in Prolog as an assignment in our first year. Granted, it did not make "objects", but what it did do was group NP's, VP's, adjectives etc etc in a tree-like structure. I am very confident that it could be rewritten to do most of the things that are claimed here. It even had (rough) support for so-called garden path sentences, and breadth/depth-first search. A lot of research is conducted here in Utrecht (The Netherlands) to have good NLP's. Ofcourse, syntax isn't the hardest part, trying to get good semantics is a lot more troublesome, although I know of more than enough people who would like to spank me for saying this.
Slashdot Mirror
Ahem. How is this news? In fact, we (AI students) had to write a NLP (Natural Language Parser) in Prolog as an assignment in our first year. Granted, it did not make "objects", but what it did do was group NP's, VP's, adjectives etc etc in a tree-like structure. I am very confident that it could be rewritten to do most of the things that are claimed here. It even had (rough) support for so-called garden path sentences, and breadth/depth-first search. A lot of research is conducted here in Utrecht (The Netherlands) to have good NLP's. Ofcourse, syntax isn't the hardest part, trying to get good semantics is a lot more troublesome, although I know of more than enough people who would like to spank me for saying this.
Sure, and how many pro audio apps are there yet? It's nice to have a good framework, but let's see it put to good use.