Re: [OmT] TM pool, was: search and replace
- There is a lot of talk about pooling TMs at the moment, but I am one of
those who is very sceptical about it.
Firstly, I don't think the copyright issue can be resolved all that
easily. Some people are presenting a scenario in which TMs for any
translation are routinely made publicly available, and I can't see that
Secondly, I think the value of TMs is seriously overrated. TMs of most
translations are not of much use to anyone, mainly because their content
isn't sufficiently similar to anything else. Then there is the quality
issue, which some people are starting to raise.
What I envisage is that the TMs of some translations will be
disproportionately valuable. A TM of all of Microsoft's UI strings is
obviously immensely valuable. So is the TM of all the EU legislation.
The same would be true of a lot of national legislation, formal
standards, and texts of this kind. In fact, a TM of standards texts
might be a way of making useful information available to the translators
who need it (like me), without them having to pay the very high prices
for standards. But this is a very specific case, and the owners of the
copyright would have to be persuaded.
If TMs really were as useful as many people seem to think, we should be
seeing far greater advances in the area of alignment tools: intelligent
alignment tools that could be used to align existing content (e.g. on
the Internet) with no manual intervention. This is not that difficult to
implement - absolutely trivial compared to machine translation, for
instance. But it still requires some investment, whereas at the moment
all the talk is about building repositories of TMX files that already
exist. This is attractive to some business people because they envisage
an opportunity to get something for nothing and make lots of money out
of it. The problem is that what you get for nothing isn't always worth
- Thanks for all your responses! Interesting to read your thoughts about
the various ins and outs. It is even more complicated than I
suspected... I imagined anonymous, generic, quite clearcut,
non-specific TMs. Entry-level starter TMs, as it were. like a simple
little dictionary. But maybe glossaries are more the way to go then as
there is no issue of sentence matching or confidentiality. Having TMs
online while you work on them seems a rather impossible proposition.
Thanks for helping me further and for the links to/mention of other sites.
On Sun, Aug 31, 2008 at 11:35 AM, Martin Wunderlich <martin_wu@...> wrote:
> Creating just such a TM repository would be one of the advantages of using a
> TM server (see my recent suggestion). If OmegaT can connect to such a TM
> server, it would be no problem to create a public repository, free for
> everyone to use (read and write). It would also fill up automatically as
> people translate away while they are connected to the TM server.
Elise Reynolds - id est
- On Mon, Sep 1, 2008 at 6:16 AM, Elise Reynolds <elirey@...> wrote:
> Thanks for all your responses! Interesting to read your thoughts aboutSee, now, a glossary repo sounds like a great idea, to me.
> the various ins and outs. It is even more complicated than I
> suspected... I imagined anonymous, generic, quite clearcut,
> non-specific TMs. Entry-level starter TMs, as it were. like a simple
> little dictionary. But maybe glossaries are more the way to go then as
> there is no issue of sentence matching or confidentiality. Having TMs
> online while you work on them seems a rather impossible proposition.
> Thanks for helping me further and for the links to/mention of other sites.
I never have the time to build glossaries.
translation and interpreting
English, French, Portuguese, Spanish