Re: Re: CAT software
- Vit Ruzicka wrote:
> Well, I have read a little about CAT lately and my opinion is that thevery idea of CAT may not be dismissed lightheartedly. Just a few remarks:
I can't recall Computer-Aided Translating (that's what CAT stands for, isn't
it - I hate abbreviations) discussed on the list, but whatever has been said
I'm sure anything that can be put in an algorithm, can be processed by a
"computer". It depends on the technology level. So why don't we just set up
a business to be the first to develop such technology so we are not the ones
who will lose by such development? I know, there ain't enough cash. Let's go
on translating the primitive tedious way, then.
- Hi everybody
I have somehow been out of touch, although receiving the daily digests.
Well, I have read a little about CAT lately and my opinion is that the very idea of CAT may not be dismissed lightheartedly. Just a few remarks:
Many years ago, computers were believed not to be able to play chess, ever. Those days, the chess game was considered too creative for computers to cope. Today, programmes like Chessmaster simply compare a database of 300 000 chess games with a single move in a game and are able to achieve a high standard of excellence in the final product, i. e. the chess game for this instance.
Mentioning the chess and computers, many similarities may be drawn to CAT. Again, a database is compared with a "single move" (a translation of a segment, which is usually a sentence) and a suggestion is made on a basis of database contents.
Just when the time comes of readily available reliable databases of millions of "segment translation solutions" nobody can really be certain.
>I'm sure anything that can be put in an algorithm, can be processed by aA pretty big leap to assume that language translation can be
>"computer". It depends on the technology level.
algorithmic, isn't it? I rather suspect it can't.
>So why don't we just set upPeople have been working on it for decades with very limited results.
>a business to be the first to develop such technology so we are not the ones
>who will lose by such development? I know, there ain't enough cash. Let's go
>on translating the primitive tedious way, then.
Right now the most promising avenue seems to be "controlled
language"--i.e. writing texts specifically to be processed by machine
translation systems. I suppose a lot of the computer manuals that
have been our bread and butter lately could theoretically be written
- Hi everybody in the New Year,
I'd like to say someting about the CAT tools as I have been studying them
and working with them for a couple of years now. I think there exists a
common misunderstanding what this acronym stands for.
CAT tools can be divided into MT and MAHT tools. The first acronym stands
for Machine Translation and there is still a long way to go, in order to
achieve really acceptable results.
The second stands for Machine Aided Human Translation. This is the kind of
tools I have been using. Some examples - Trados Translator's Workbench, IBM
Translation Manager, Atril's DejaVu, which is my favorite. These tools are
really valuable if you have repetitive translations as software
localiyations, technical manuals, etc. This type of tools is quite mature
and of a great help to any experienced technical translator.
MT tools are focused on people having little or none language knowledge and
they are, at least nowadays, able to give you just a very high-level
overview of a document contents, sometimes in a funny language. However, Vit
is right these tools are constantly being improved and with the increasing
computing power the results are also beginning to be more and more
acceptable. But the way to a perfect translation is still long and, by my
opinion, unachievable, due the each language variability. More complicated
texts will remain in hands of professional translators, the easier ones will
be automatically processed using MT tools.
> >I'm sure anything that can be put in an algorithm, can be processed by aI leave this to everybody to decide for herself or himself.
> >"computer". It depends on the technology level.
> A pretty big leap to assume that language translation can be
> algorithmic, isn't it? I rather suspect it can't.