Re: Machine translation. Was: Re: Translation platform migration: status update

Benson Muite <[email protected]>
Newsgroups gmane.linux.redhat.internationalization
Message-ID <[email protected]>
On 8/18/19 9:56 PM, Piotr Drąg wrote:
> niedz., 18 sie 2019 o 20:48 Jean-Baptiste <[email protected]>
> napisał(a):
>> As all we do is open source, it's fine to reuse our work for any activity as long as it respect the license.
>>
>> If such activities interests you, here is an extraction of all our translation I did for the amagama team:
>>
>> https://jibecfed.fedorapeople.org/partage/2019-07-06-all-fedora-zanata-org-tm.7z
>>
>> And here is the script to run it yourself:
>> https://github.com/Jibec/fedora-translation-statistics/blob/master/get_zanata_tm.py
>>
>> You just have to change the headers;
>> headers = {"X-Auth-User":"jibecfed", "X-Auth-Token":"change-me"}
>>
>> Note: it produce one multilingual translation memory file (tmx) per project-iteration. If you had a Python script to transform+merge it in one tmx per language with all projects, it would be great!
It seems possible to configure Weblate to do this for all projects in a 
single Weblate instance. Ideally would want to do this across a larger 
number of translation projects for open source software.
>>
>> Would there be community objections to use of the translations as
>> machine translation training material?  If not, this might allow for
>> some collaboration with machine translation companies in a similar way
>> to the use of reCaptcha for crowdsourcing image labels.
> I’m pretty sure translators have copyright over their work and you
> can’t reuse translations without attribution (additionally to any
> license concerns.) I’m not a lawyer, of course.
>
> Best regards,
Ok.  License terms for use of translation memory may need some thought - 
if it is an exact string match of a previous translation, does 
attribution go to the person that first entered the translation or to 
the person that approved the translation?  If it is not an exact match, 
and one has a derived work, what more should be done? At present Pontoon 
assigns credit to the person that authorizes the translation, even if it 
came from translation memory suggestions or a machine learned 
suggestion. Pontoon's terms and conditions do not seem to be very 
specific in this area - https://pontoon.mozilla.org/terms/
_______________________________________________
trans mailing list -- [email protected]
To unsubscribe send an email to [email protected]
Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/
List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines
List Archives: https://lists.fedoraproject.org/archives/list/[email protected]
lmpx.com only provides a reader for public news (NNTP) servers. It is not affiliated with the servers or forums shown here and is not responsible for the content of articles, which is written by their respective authors.