I was reading in the LISA newsletter about a form of translation memory software that is not tied to the sentence but instead to large paragraphs and that creates a big index of all elements of a text.
The article is by Timothy Hunt of TermSeek Inc. and actually comes from the archives but I havent seen it before.
The software called translation support software and two kinds being worked on are TransSearch and Translator’s Intuition:
bq. For example, Elliott Macklovitch and Graham Russell in What’s been Forgotten in Translation Memory pointed out that statistical based TM systems would say that sentence (1) below is closest to sentence (2) even though sentence (3) is closer in meaning.
bq. The wild child is destroying his new toy.
The wild chief is destroying his new tool.
The wild children are destroying their new toy.
There is also a list of language and translation blogs (under the subdirectory humor).