Re: [OmT] The need to regularly save the translation memory
- On 6/1/05, Samuel Murray <leuce@...> wrote:
> > Thanks, Maxym. I find the idea of OmegaT quite interesting, but I amThanks, Samuel and Maxym. I was aware that OmegaT does not translate
> > bit skeptical about having it, say, translating 80% of a text after
> > some time. Am I wrong?
> If you're talking about fuzzy matching across the entire text, then 80% is huuuuuuuuge. I recently had a text which had a 40% direct match and a futher 20% high fuzzy match across the entire text, but that is extremely rare. You'll only find the 80% scenario if you're translating a manual or suchlike which has been translated before and in which only a few sentences were changed.
> Besides, since OmegaT segments by paragraph, it is highly unlikely that you'll find a situation within a single text or subject field where there is a definite need to translate two instances of a paragraph differently.
in the strict meaning of the word translate and I was speculating
about the potential percentage of matches.
- Paul Smith skryf om 11:55 AM op 01/06/2005:
> Thanks, Samuel and Maxym. I was aware that OmegaT does not translateIn that case, well if the match is 80% then it means that OmT has "translated" 80% of that particular paragraph, but you still need to check the entire paragraph.
> in the strict meaning of the word translate and I was speculating
> about the potential percentage of matches.
Take the sentence "[demonstrative] [verb to-be] [object]", which in Langauge X may consist of three words. If the object changes from singular to plural, OmT will mark it as a 66% match because 2 out of 3 words are still the same, but in English you'd have to change not one but two words (the verb has to agree with the noun), and in some other languages you may have to change all three words (the demonstrative has to agree with the verb also).