Date: prev next · Thread: first prev next last
2014 Archives by date, by thread · List index


Hi :)
Yes, definitely only as a helper NOT as a replacement!  However, looking
back through the archives suggests that Michael is not the only one who
finds that trying to use an MT actually slows him down and is more of a
hindrance than a help.

I've not encountered TM before, i will have to look it up.

Errr, i thought Pootle and other things were more helpful in having some
sort of database where humans could look-up phrases and see how other
humans had translated them.  Still lacks context so nuances and such could
still be dodgy.  Is that how "fuzzy string"s appear?

If Pootle doesn't have some functionality switched on then it's probably
for a very good reason and has probably been discussed and decided before.
I haven't checked the archives and probably wont have time to either,
sorry! :(


The main response i seem to be getting is that MTs are only good in a very
limited way and probably only for very occasional use but it's really down
to individual translators and which languages are being translated to or
from.

Thanks and regards from
Tom :)



On 3 November 2014 07:26, Rimas Kudelis <rq@akl.lt> wrote:

Hi,
Let me chime in here as well. First of all, Pootle has the ability to use
a configured TM. Couple options available out of the box are Google
Translate (which is an MT, not a TM service) and Aperture (which, if I
remember correctly, holds a collection of translated strings from a large
amount of open source software, and can suggest these translations to you).
In addition to this, there's apparently a lookup feature as well, with an
example config for Wikipedia. I'm not sure why we aren't using any of these
at the moment, but I guess we could turn them on, if needed.

However, none of these features are even close to being a replacement for
a real human localizer being at the moment. They are mere helpers, and
nothing more. Surely, the level of accuracy varies by language, but
considering other replies, I guess even translations into English are bad
enough. As a user, I sometimes stumble upon software, which is "localized"
into Lithuanian using Google Translate (or something very similar). Tell
you what: on these cases I feel insulted by the idiocy of that app's
"localizer", because that person obviously didn't know at all what they
were doing and how shitty the outcome of their "effort" is. I can't tell
for sure that that outcome isn't helpful, but I really doubt that it is.
So, to summarize my point, we should never attempt to seed any locale with
machine translations. These are a good helper mechanism, but a bad base to
start building up on.

By the way, writing this also reminded me of Google's Code-In programme
and what we should learn from it. GCI is similar to Google Summer Of Code,
but targets younger students to whom the participating projects assign
smaller than GSOC tasks. Among them, localization tasks were also
acceptable. So, some students quickly found a way to easily cash out the
rewards offered for completing these tasks, by using MT services. What's
worth learning from this is that there should always be a mentor within a
project who understands the target language well enough to at least tell
whether it looks like a result of MT. Otherwise it's just a waste of
rewards and its reputation.

Regards,
Rimas

On 2014 m. lapkritis 3 d. 00:03:21 EET, Tom Davies <tomcecf@gmail.com>
wrote:
Hi :)

Going off-topic (now that the original question has been solved),
hopefully
just briefly ...

Of course Firefox (and many others) allows add-ons such as
machine-translators.  They are getting much better but are still
hilariously bad quite often.  Often they give just enough of a hint
that i
think i understand what someone is saying although i always wonder if
they
have sent me off in a wrong direction.

My MT (i think "Quick translate"), gave me;
"Hello!
I ask you to add the Abkhaz language for translation
Libre Office ver. 4.3.3.2
"
Which made a lot of sense and it's good to have a respectable human
confirm
that because it gives me a little more confidence in the MT.


I've often wondered if they might be "good enough" to get rough
translations done well enough for humans to proof-read and polish?
Perhaps
just "good enough" to use alongside the human translators own skill and
knowledge, perhaps to get some inspiration?  Perhaps better for people
who
are only just starting to translate things?

I've also wondered if it's easier to have paired teams.  So in this
case
someone who is a native Russian-speaker but understands Abkhazian "well
enough" to do first drafts and then a native Abkhazian-speaker to do
proof-reading, ideally one who understands Russian (or English or
something) just well enough to be able to look back at a source
document to
double-check that things haven't gone too far off-track.

Does either of those ideas have any validity?  Are they something that
noobs or laymen often seem to think but turns out to be more work
and/or
less accurate than whatever different ways your teams use?

Err, i am a typical English person and only understand 1 language at
all
and not even that great at that 1 so please forgive my noobishness in
this
post.
Regards from
Tom :)




On 2 November 2014 19:17, Sophie Gautier <gautier.sophie@gmail.com>
wrote:

Hi

Le 2 nov. 2014 20:10, "Andras Timar" <timar74@gmail.com> a écrit :

On Sun, Nov 2, 2014 at 7:52 PM, Sophie <gautier.sophie@gmail.com>
wrote:
Hi and welcome,
Le 02/11/2014 17:30, Андрей Абухба a écrit :
 Здравствуйте!

Прошу Вас добавить абхазский язык для перевода
Libre Office ver. 4.3.3.2

Could you write in English? it will be much more easier for us to
help
you :)
Thanks
Sophie


I think Andrey wanted us to add Abkhazian language to Pootle. I've
just done that.(Probably he cannot write in English, and will
translate LibreOffice from Russian to Abkhazian, but I don't know
for
sure.) Anyway, welcome Andrey, and you can start translating
LibreOffice 4.3 UI in Pootle.

Ha great! Thanks Andras. Andrey let us know which language is
preferred for
you so we know who to ping to help you if you need it.
Cheers
Sophie

--
To unsubscribe e-mail to: l10n+unsubscribe@global.libreoffice.org
Problems?
http://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guidelines + more:
http://wiki.documentfoundation.org/Netiquette
List archive: http://listarchives.libreoffice.org/global/l10n/
All messages sent to this list will be publicly archived and cannot
be
deleted


--
To unsubscribe e-mail to: l10n+unsubscribe@global.libreoffice.org
Problems?
http://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guidelines + more:
http://wiki.documentfoundation.org/Netiquette
List archive: http://listarchives.libreoffice.org/global/l10n/
All messages sent to this list will be publicly archived and cannot be
deleted

--
Rimas

--
To unsubscribe e-mail to: l10n+unsubscribe@global.libreoffice.org
Problems?
http://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guidelines + more: http://wiki.documentfoundation.org/Netiquette
List archive: http://listarchives.libreoffice.org/global/l10n/
All messages sent to this list will be publicly archived and cannot be
deleted


-- 
To unsubscribe e-mail to: l10n+unsubscribe@global.libreoffice.org
Problems? http://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guidelines + more: http://wiki.documentfoundation.org/Netiquette
List archive: http://listarchives.libreoffice.org/global/l10n/
All messages sent to this list will be publicly archived and cannot be deleted

Context


Privacy Policy | Impressum (Legal Info) | Copyright information: Unless otherwise specified, all text and images on this website are licensed under the Creative Commons Attribution-Share Alike 3.0 License. This does not include the source code of LibreOffice, which is licensed under the Mozilla Public License (MPLv2). "LibreOffice" and "The Document Foundation" are registered trademarks of their corresponding registered owners or are in actual use as trademarks in one or more countries. Their respective logos and icons are also subject to international copyright laws. Use thereof is explained in our trademark policy.