I also entirely fail to believe that a lexicon of dying tongues can be constructed over the Internet. The notion of tribesmen with PCs and `net connections brings a dreamlike smirk to my face. I think they call it FIELD linguistics for a reason.
It seems to me that this project has overlooked two tremendous stumbling blocks. The first involves context/ambiguity. Take the English, "it's pretty bad outside" Now, for an English speaker, this is no trouble, since the "it's" is generally held to be referring to the weather. Other languages lack such a frame of reference. Secondly, I believe that the "distributed" property of the system leaves it widely open to poor or intentionally incorrect translations, unless the system is employing some statistical method for finding the "mean translation" of a phrase out of a batch of candidates. While I appreciate this researcher's work on Machine Translation, I think that this might better be served by designing some type of meta-language with a superset of linguistic features from which native translations might be compiled.
Slashdot Mirror
Syntactic translation is mentioned nowhere in the article. This researcher is attempting to build a LEXICON.
I also entirely fail to believe that a lexicon of dying tongues can be constructed over the Internet. The notion of tribesmen with PCs and `net connections brings a dreamlike smirk to my face. I think they call it FIELD linguistics for a reason.
It seems to me that this project has overlooked two tremendous stumbling blocks. The first involves context/ambiguity. Take the English, "it's pretty bad outside" Now, for an English speaker, this is no trouble, since the "it's" is generally held to be referring to the weather. Other languages lack such a frame of reference. Secondly, I believe that the "distributed" property of the system leaves it widely open to poor or intentionally incorrect translations, unless the system is employing some statistical method for finding the "mean translation" of a phrase out of a batch of candidates. While I appreciate this researcher's work on Machine Translation, I think that this might better be served by designing some type of meta-language with a superset of linguistic features from which native translations might be compiled.
Exactly. Now, where do we get these textures from? Does an X server keep some sort of cache that could be shuffled off into texture memory?
Does this mean I'll finally get to use a version of enlightenment that lets me push and pull windows further and closer to me?