Really 8,000 words more for Master%

Hi,

I've just seen that there are suddenly app. 8,000 words more for Master in Upper Sorbian. Now there are 109,283 words, previously 101,873. Lower Sorbian shows less words now, 99,243, previously 101,028, I think. What has happened here? Did you mix Upper Sorbian and Lower Sorbian? Why do these languages have a different number of English words?

Michael

Hi Michael, *,

Pootle updates on master are still work in progress (read: ongoing).

Keep cool and relaxed. :wink:

Michael Wolf <milupo@sorbzilla.de> hat am 12. April 2017 um 21:31 geschrieben:

Hi,

I've just seen that there are suddenly app. 8,000 words more for Master
in Upper Sorbian. Now there are 109,283 words, previously 101,873. Lower
Sorbian shows less words now, 99,243, previously 101,028, I think. What
has happened here? Did you mix Upper Sorbian and Lower Sorbian? Why do
these languages have a different number of English words?

rgds
Sophia Schroeder

Sophia Schroeder schrieb:

Hi Michael, *,

Pootle updates on master are still work in progress (read: ongoing).

Keep cool and relaxed. :wink:

Hi Sophia,

what does it mean? Are the stats reliable or not? Can I keep translating or, could it be that the translation of the new strings will be a waste of time?

Michael

Hi Michael, *

as cloph already mentioned, sources of some strings had been changed.

Some of them are already translated somewhere and are still in the translation memory system.

Unfortunately there is no automatic tool (yet) to put these given translations into the new strings and be done.

Yes it is worth to translate even further. :wink:

    Michael Wolf <milupo@sorbzilla.de> hat am 13. April 2017 um 17:44 geschrieben:

    Sophia Schroeder schrieb:

        > >
> Hi Michael, *,
>
> Pootle updates on master are still work in progress (read: ongoing).
>
> Keep cool and relaxed. :wink:
>
> >
    Hi Sophia,

    what does it mean? Are the stats reliable or not? Can I keep translating
    or, could it be that the translation of the new strings will be a waste
    of time?

    Michael

    --
    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

Regards / Mit freundlichen Grüßen
Sophia Schroeder

Hi Michael, *,

Hi,

I've just seen that there are suddenly app. 8,000 words more for Master in
Upper Sorbian. Now there are 109,283 words, previously 101,873.

The increase in total works is fishy, pootle did not remove strings
that were removed from files.

Lower
Sorbian shows less words now, 99,243, previously 101,028, I think. What has
happened here? Did you mix Upper Sorbian and Lower Sorbian? Why do these
languages have a different number of English words?

I assume only translated strings are affected, i.e. translated
strings, but removed from templates → still listed in pootle.

e.g https://translations.documentfoundation.org/de/libo_ui/translate/starmath/source.po#unit=79252016
→ the Color Fuchsia was removed (and is removed from the files on
disk, there it is an obsolete string for de)

Strings not kept for Friulian on the other hand that did not have this
translated.

It is also not that it would add *all* obsolete strings, as in case
for german there are many more obsolete strings that were not
(re-)added.

So there's a bug in removing translations from webinterface/database.

ciao
Christian

Hi Michael,

Sophia Schroeder schrieb:

Hi Michael, *,

Pootle updates on master are still work in progress (read: ongoing).

Keep cool and relaxed. :wink:

Hi Sophia,

what does it mean? Are the stats reliable or not?

Stats in terms of total wordcount are unreliable, in terms of
untranslated it should be OK (at least from first glance, see last
mail, seems only already translated strings were not removed properly)

Can I keep translating or,
could it be that the translation of the new strings will be a waste of time?

There *might* be some autotranslations of the strings with only
context-changes, but don't hold your breath for that.

ciao
Christian

This is getting absurd. I had just about gotten the last batch of updates down to about 2k, most of which were I-don't-know-what changes which left a high number of translated strings fuzzy. Now Gaelic is back up to 28k and clicking through them, it looks like almost all are 100% TM matches somewhere.

How is this even vaguely respectful of translator time?

Right now, I'm not so sure that it's worth my time translating further cause I'm not translating, am I? I'm a human TM-match-insertion-tool.

Michael

Sgrìobh Sophia Schroeder na leanas 13/04/2017 aig 17:04:

Hi Michael, *

I am the wrong person to give you proper answers. Am just a translator like you.

I just wrote from my personal POV and vague knowledge from what I'm getting.

Christian Lohmaier is the right reference person for such things.

So look at his mails about this issues. :slight_smile:

    Michael Bauer <fios@akerbeltz.org> hat am 13. April 2017 um 19:34 geschrieben:

    This is getting absurd. I had just about gotten the last batch of
    updates down to about 2k, most of which were I-don't-know-what changes
    which left a high number of translated strings fuzzy. Now Gaelic is back
    up to 28k and clicking through them, it looks like almost all are 100%
    TM matches somewhere.

    How is this even vaguely respectful of translator time?

    Right now, I'm not so sure that it's worth my time translating further
    cause I'm not translating, am I? I'm a human TM-match-insertion-tool.

    Michael

    Sgrìobh Sophia Schroeder na leanas 13/04/2017 aig 17:04:

        > >
> Some of them are already translated somewhere and are still in the translation memory system.
>
> Unfortunately there is no automatic tool (yet) to put these given translations into the new strings and be done.
>
> Yes it is worth to translate even further.:wink:
>
> >
    --
    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

Regards / Mit freundlichen Grüßen
Sophia Schroeder

Hi Sophie,

I know, sorry if you felt this was aimed at you - it wasn't - though I am quite certain your assessment was correct, which is why I "replied" to your post. It was aimed at whoever is letting this happen again and again and again. This has been a complaint for as long as I can remember on LO and I'm somewhere between gutted and furious that it's still happening. Sure, maybe a locale like German or French has the translators to throw at another 28k of work that was already done but most locales aren't that fortunate. Take Uyghur, it was at 97% when OO and LO split. There hasn't been much translation activity in this locale in both OO and LO since and there haven't been *that* many genuinely new strings in LO actually but that has left Uyghur on 61% and if I click through the strings on Uyghur, it's one 100% match after the other.

Yes, this is an angry post (not angry at Sophia) but before someone slaps me down for being angry, maybe ask yourself at which point getting angry is justified in a FOSS project when someone somewhere for some arcane reason keeps wasting your lifetime as if it didn't matter. Do we really want to burn out localizers like they were a dime a dozen?

Michael

Sgrìobh Sophia Schroeder na leanas 13/04/2017 aig 19:05:

Hi Michael,
I'm in the middle of a meeting, so making it short for now to correct one
thing. Until last year I was alone to maintain the FR translation and now
we are 2. So the pain is the same for quite everybody here. I spend nights
on it because my government, large and small companies, association etc.
are depending on it.
We are investigating with Cloph where the problem is lying. I will keep you
updated later tonight when I'll be back home.
Cheers
Sophie

I am Sophia(S), not Sophie Gautier,

I hope you don't confuse us two. :wink:

And I am relatively new and maybe overenthusiastic here with my contribution part.
I was lurking around OOo / LO since beginning.

Michael Bauer <fios@akerbeltz.org> hat am 13. April 2017 um 20:21 geschrieben:

    Hi Sophie,

    I know, sorry if you felt this was aimed at you - it wasn't - though I am quite certain your assessment was correct, which is why I "replied" to your post. It was aimed at whoever is letting this happen again and again and again. This has been a complaint for as long as I can remember on LO and I'm somewhere between gutted and furious that it's still happening. Sure, maybe a locale like German or French has the translators to throw at another 28k of work that was already done but most locales aren't that fortunate. Take Uyghur, it was at 97% when OO and LO split. There hasn't been much translation activity in this locale in both OO and LO since and there haven't been *that* many genuinely new strings in LO actually but that has left Uyghur on 61% and if I click through the strings on Uyghur, it's one 100% match after the other.

    Yes, this is an angry post (not angry at Sophia) but before someone slaps me down for being angry, maybe ask yourself at which point getting angry is justified in a FOSS project when someone somewhere for some arcane reason keeps wasting your lifetime as if it didn't matter. Do we really want to burn out localizers like they were a dime a dozen?

    Michael

    Sgrìobh Sophia Schroeder na leanas 13/04/2017 aig 19:05:

        > >
> Hi Michael, *
>
> I am the wrong person to give you proper answers. Am just a translator like you.
>
> I just wrote from my personal POV and vague knowledge from what I'm getting.
>
> Christian Lohmaier is the right reference person for such things.
>
> So look at his mails about this issues. :slight_smile:
>
> > --
    Akerbeltz http://www.faclair.com/
    Goireasan Gàidhlig air an lìon
    Fòn: +44-141-946 4437
    Facs: +44-141-945 2701

    Tha Gàidhlig aig a' choimpiutair agad, siuthad, feuch e!
    Iomadh rud eadar prògraman oifis, brabhsairean, predictive texting,
    geamannan is mòran a bharrachd. Tadhail oirnn aig www.iGàidhlig.net http://www.iGaidhlig.net/

Regards / Mit freundlichen Grüßen
Sophia Schroeder

Hi all,

BTW when I try to upload large file, I always get 504 gateway timeout error,
I used formula/source/core/resource.po file in Kazakh (kk) version.
In fact this PO does not require translation, I just copied the
messages to translations,
since those are all formulas.
But now I am unable to upload it to Pootle,

Hi All,

I'm jumping on behalf of the Pootle developers, as we seem to have multiple
threads and issues happening at once, and not all of them actually relate
to Pootle:

1. Everything going fuzzy/untranslated because of context change. I agree
I wouldn't like this as a translator especially if these are just string
moves. There are ways to script such things in the PO files to prevent
them being loaded as fuzzy. We're reaching out to the TDF l10n team to see
how we can help in future and to see if we can't clean this up for the
current change.

2. Some strings being removed from PO but still showing on Pootle, we're
investigating how this might have happened. We've done extensive testing
in this area so we're surprised this is happening but will look at the
factors. TDF server is large so it does take some time to update.

3. Uploads. Please let us know when these happen, it often it a matter of
tuning the upload size and timeout but hearing after the fact doesn't help
us. Let us know straight away please. At the moment I expect uploads to
be difficult because of the initialisation of new PO files on the server.
Waiting till that is complete will likely be a better option.

Trust that helps give some clarity. And I trust give some clarity as to
where Pootle is involved in this process.