Re: [OmTdev] Architecture Ideas.

[Wayne W. <wa...@wa...>]

Samuel,

That's not absolutly necessary. Have you looked at WordFasts VLTM?
It's exactly that but based on sentences.

It explains you can use the dictionary without contributing yours
for confidentiality or secrecy reasons. But if you don't have a
secrecy requirements on a particular assigment, you can click
and submit your segments.

That seems smart. So the decision to contribute can be done per
project rather than reviewing monotonously through terms in a =
dictionary.

STILL, the benefits of building up your own dictionary over time
are huge. Consider the number of repeated terms throughout the
document.=20

Wayne

-----Original Message-----
From: ome...@li...
[mailto:ome...@li...] On Behalf Of
Samuel Murray
Sent: Tuesday, September 18, 2007 12:24 PM
To: Mailing list for OmegaT developers.
Subject: Re: [OmTdev] Architecture Ideas.

Wayne Walter wrote:

> If the tool is really good and easy to use, has all the features,
> we can get other translators to pool together so the dictionary grows
> faster.

What you're proposing is a dictionary creation project.  Before a=20
translator can contribute his dictionary pairs, he would have to review=20
them to ensure that there is no breach of confidentiality.  This=20
reviewing takes time, and I wonder how many translators would be willing =

to contribute their time for something they are unlikely to benefit=20
from, unless there are thousands of participants.

> ...it would be easier=20
> To get more translators to join the project and build a common GPL
licensed
> Dictionaries.

I would not want to contribute to a GPL dictionary.  GPL is a terrible=20
licence that robs people of several freedoms.  I suppose other people=20
might have other objections to participate in such a project.  I know of =

a guy who refuses to release his dictionary as any open source licence=20
because it would mean that he would lose the ability to ensure high=20
quality... but anyone can still get and use his dictionary free of=20
charge for any purpose (including commercial).

> Once you use a couple of keys to tell it clavier goes with keyboard
> and =E9cran goes with screen. Then it will have these terms in the =
future.

This is a wonder thing to have, if you can write it, but I wouldn't use=20
it to create a dictionary.  Instead, I would use it to create a=20
glossary.  Semi-automatic terminology creation is a wonderful idea.  I=20
believe linguists refer to it as "word alignment" or "bilingual term=20
extraction".  If you can build such a system in Java, it would be=20
extremely helpful because currently only SDL Multiterm Extract offers=20
remotely the same functionality, at EUR 500.00 per single-user licence.

But that would be a separate thing from OmegaT.  While you're at it you=20
might also design an aligner that is more user-friendly than Bligner.=20
But again, while these tools are useful for a translator who does some=20
pre-work before he starts translating, I see little benefit in having it =

included as a process in OmegaT.

> The translator can match them up and even select if they are a =
glossary
> term, etc.
>=20
> The power really grows, of course, if more translators join the =
project
> and contribute dictionaries and glossaries.

Dictionary creation tools that use bitext as sources would be wonderful, =

especially for minority languages that do not have large open source=20
dictionaries available.  But this has little to do with OmegaT.  It is a =

separate project that can involve volunteers who feel strongly about a=20
language but who are not necessarily translators.  If you're seriously=20
considering writing such tools, I'd love to discuss it with you because=20
I'm involved in the upliftment of minority languages myself, and there=20
is a dire need for the type of tools you describe.

Samuel


-------------------------------------------------------------------------=

This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2005.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
_______________________________________________
Omegat-development mailing list
Ome...@li...
https://lists.sourceforge.net/lists/listinfo/omegat-development