Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss
Hi all
9 views
Skip to first unread message
Ibrahima SARR
Mar 25, 2015, 4:44:17 PM3/25/15
to Firefox l10n, Dwayne Bailey, dev-l10n
As you know, Firefox OS will soon be deployed in many countries in Africa.
As Fulah localization lead I would like to point out the need to expand
Fulah localizations to more local dialects. Actually, I have worked hard to
create common IT terminology of all of us but now we can create more
specific localization to suit the various regional dialects.
We can use country code along with language code to make it simple as these
usually correspond to regional variations. Also these dialects have a 3
letter code if necessary.
Senegal: ff_SN (same for Mauritania, Gambia)
Guinea: ff_GN (same as Burkina Faso, Mali, Sierra Leone, Guinea Bissau)
Nigeria ff_NG (same as Cameroon, Chad, Niger, Central Africa)
These are not 100% identical but they will diversify the localizations
enough to include most dialects.
Si it would be interesting to see which of these can be created in Pootle.
I thnik Mali and Guinea are more of a priority now and share the same
diaclect with Guinea Bisau, so that could be created under one folder.
Then I will help get/train localizers.
How feasible and useful is that?
Cheers
Ibrahima Sarr
As Fulah localization lead I would like to point out the need to expand
Fulah localizations to more local dialects. Actually, I have worked hard to
create common IT terminology of all of us but now we can create more
specific localization to suit the various regional dialects.
We can use country code along with language code to make it simple as these
usually correspond to regional variations. Also these dialects have a 3
letter code if necessary.
Senegal: ff_SN (same for Mauritania, Gambia)
Guinea: ff_GN (same as Burkina Faso, Mali, Sierra Leone, Guinea Bissau)
Nigeria ff_NG (same as Cameroon, Chad, Niger, Central Africa)
These are not 100% identical but they will diversify the localizations
enough to include most dialects.
Si it would be interesting to see which of these can be created in Pootle.
I thnik Mali and Guinea are more of a priority now and share the same
diaclect with Guinea Bisau, so that could be created under one folder.
Then I will help get/train localizers.
How feasible and useful is that?
Cheers
Ibrahima Sarr
Dwayne Bailey
Mar 26, 2015, 5:26:57 PM3/26/15
to Ibrahima SARR, dev-l10n, Firefox l10n
>From a Pootle perspective, a lot is possible.
But I think the critical questions would be if the effort creates the
needed value. Remember that Arabic which has many regional variants is
usually only translated in one form and this is a language with 400 million
speakers. I don't really know how far apart the dialects are, but for each
dialect you are essentially creating another new translation which need to
be maintained. Does the Fulah team have the resources to maintain all
three?
--
Dwayne
*Translate*
+27 12 460 1095 (work)
But I think the critical questions would be if the effort creates the
needed value. Remember that Arabic which has many regional variants is
usually only translated in one form and this is a language with 400 million
speakers. I don't really know how far apart the dialects are, but for each
dialect you are essentially creating another new translation which need to
be maintained. Does the Fulah team have the resources to maintain all
three?
Dwayne
*Translate*
+27 12 460 1095 (work)
Axel Hecht
Mar 27, 2015, 7:48:03 AM3/27/15
to Dwayne Bailey, Ibrahima SARR, Firefox l10n
I'd second what Dwayne said, if we end up having three-times the work,
we'll see only little returns.
To ask a different way, are there systematic approaches one can take to
support variants? Like, en-GB uses tooling to catch the strings that are
different to en-US, and IIRC, also creates them for the most part. en-ZA
OTH doesn't have that tooling, and is often lagging behind. In its case,
it falls back to en-US, which is only half-bad, but for a variant of
Fulah, that'd be ughly.
If such algorithms exist, then you could do "Fulah", and then in a
second step do high-quality variations of it. Not sure if pootle would
be the best tool to do this. Something closer to the metal might be better.
Axel
we'll see only little returns.
To ask a different way, are there systematic approaches one can take to
support variants? Like, en-GB uses tooling to catch the strings that are
different to en-US, and IIRC, also creates them for the most part. en-ZA
OTH doesn't have that tooling, and is often lagging behind. In its case,
it falls back to en-US, which is only half-bad, but for a variant of
Fulah, that'd be ughly.
If such algorithms exist, then you could do "Fulah", and then in a
second step do high-quality variations of it. Not sure if pootle would
be the best tool to do this. Something closer to the metal might be better.
Axel
0 new messages