I was testing Yahoo's keyword extraction service again today.
Well, I've been thinking about building a multilingual news site or search system that shows articles on a same topic in more than two languages so that the readers can compare how the same incident is described by the writers from different countries.
I deemed this is not so tough and I could do that by combining some web APIs, but soon after I really tried to make it, I've gotten stuck.
The problem is how to find equivalent words in different languages. There are bunch of nice online dictionaries, but to achieve smart results, the system need to choose a definition best matches with the context. Proper nouns are another problem. How can a computer program know the spelling of your favorite British band in Japanese?
Wikipedia and some translation services may have a hint. Anyways, there seems to be a long way to go...
Well, I've been thinking about building a multilingual news site or search system that shows articles on a same topic in more than two languages so that the readers can compare how the same incident is described by the writers from different countries.
I deemed this is not so tough and I could do that by combining some web APIs, but soon after I really tried to make it, I've gotten stuck.
The problem is how to find equivalent words in different languages. There are bunch of nice online dictionaries, but to achieve smart results, the system need to choose a definition best matches with the context. Proper nouns are another problem. How can a computer program know the spelling of your favorite British band in Japanese?
Wikipedia and some translation services may have a hint. Anyways, there seems to be a long way to go...
Leave a comment