Part_title tags in 5.1 help - what should l10n teams do?

l10n
#1

Hi,

with 5.1 help strings I see changes like this (this one is
hd_id3156150 from helpcontent2/source/text/writer.po):
old version in 5.0:
"<link href="text/shared/01/02220000.xhp" name="ImageMap">ImageMap</link>"
in 5.1 it gets changed to:
"<link href="text/shared/01/02220000.xhp"
name="ImageMap"><embedvar_href="text/shared/01/02220000.xhp#imagemap_part_title"/></link>"

In another help text string (hd_id3150502 in
/helpcontent2/source/text/shared/01.po) "ImageMap Editor"
got changed with 5.1 into:
"<variable id=\"imagemap_part_title\">ImageMap</variable> Editor">"

I guess the idea is that with the compilation of help files and the
build process of LO the second "ImageMap" string will get mapped into
the first string. Looks great for English source strings.

But ... some if not most languages, unlike English, have inflections
and/or do not use all-title-words capitalization, which makes this
help text programming for most of languages unusable and even
disrupting.

So, for such languages, what should l10n teams do when re-translating
these updated strings (again this is causing unneeded work for
translation teams, but I really do not want to go back there, I guess
l10n teams in this open-source project are doomed with it)?

Do we leave the "ImageMap Editor" in the first string translated as it
was in 5.0 (with just words and no additional xml tags), and also
leave the translations of strings where this part_title strings get
used as it was - without those part_titles, with plain translated
keywords? Or should we wait with localization as these strings will be
returned to their former state (and actually no new localization
effort is required)?

Thanks, m.

#2

This is a good finding. But capitalization of letters is not the biggest
problem. The paragraph “ImageMap Edite” cannot be translated correctly in
Slavonic languages (Russian and others), because they have different case-
endings. For example, “ImageMap Edite” is translated into Russian as «Редактор
Сенсорного изображения», but “ImageMap” is translated as «Сенсорное
изображение». This situation can be avoided, if the word “Edite” is deleted
from the article title.
In my opinion, the title “ImageMap Edite” for this article is good looking and
reflects the article content. So, I offer to delete <embedvar
href="text/shared/01/02220000.xhp#imagemap_part_title"/> from files
source/text/swriter/main0102.xhp, source/text/simpress/main0102.xhp and
source/text/scalc/main0102.xhp and replace it with “ImageMap”.

Best regards,
Lera

#3

I think it was an artifact from the new help editor extension. Please
always report issues like this, when you find them. I fixed this one
in master branch.

Best regards,
Andras

#4

Fortunately there are 18 more occurrences only. I'll get rid of this
crazyness now.

Regards,
Andras

#5

Hi Jay,

obviously you should learn more about natural languages.

Please read this:
https://en.wikipedia.org/wiki/Inflection
https://en.wikipedia.org/wiki/Plural
https://en.wikipedia.org/wiki/Dual_(grammatical_number)
etc.

I can give you a simple example for Slovenian:
ImageMap Editor = Urejevalnik slik s povezavami
ImageMap = Slika s povezavami

Do you get it now? Placeholder for "ImageMap" in Slovenian cannot be
used. End of story.

Not to mention there are far more complex languages in the l10n pool
of LO, so programming sentences is really out of the question for
LibreOffice.

Please undo all your changes in the code/pot files.

Thanks,
Martin

#6

I see your point. But it seems not correct. In the title 'ImageMap Editor' the
Russian phrase for Imagemap has one form (case ending), while when it goes
without Editor, just ImageMap — then it will be in another form. See the
difference below:

Сенсорного изображения: Imagemap in Russian for ImageMap Editor
Сенсорное изображение: Imagemap in Russian for just ImageMap

So there should be two different places for that phrase in Russian language.

Best regards,
Lera

#7

Hi Lera and all,

Yes these have been my changes in order to reduce duplication of translations, which is what <embed> and <embedvar> is used for.

With that particular example, we had 'ImageMap' as regular text in all module main0102.xhp files, which meant that 'ImageMap' was being translated 4 or more times, as the help page title is 'ImageMap Editor'. So using <embedvar>, i eliminated having this duplication.

I dont see an issue that translators would need to place the translated word for ImageMap in the necessary position. Example in Russian:

"Редактор <variable id=\"imagemap_part_title\">Сенсорного изображения</variable>"

Yousuf 'Jay' Philips

#8

In general, it is a very bad practice to split translatable strings into
different parts. Even if it duplicating parts of the strings increases
the work for translators, it is often the only way to gt proper
translation for many languages. The increased effort can always be
alleviated by good translation memory.

Regards,
Khaled

#9

Hi Jay, all,

Hi Jay,

obviously you should learn more about natural languages.

Please read this:
https://en.wikipedia.org/wiki/Inflection
https://en.wikipedia.org/wiki/Plural
https://en.wikipedia.org/wiki/Dual_(grammatical_number)
etc.

I can give you a simple example for Slovenian:
ImageMap Editor = Urejevalnik slik s povezavami
ImageMap = Slika s povezavami

Do you get it now? Placeholder for "ImageMap" in Slovenian cannot be
used. End of story.

Not to mention there are far more complex languages in the l10n pool
of LO, so programming sentences is really out of the question for
LibreOffice.

Please undo all your changes in the code/pot files.

Yes, and please next time, ask here before doing any change,
localization is a complex work and you have very experimented people here.

Cheers
Sophie

#10

Thanks everyone for your input into the issue. I have correct the issue by having the variable on its own separate line in the help files.

https://gerrit.libreoffice.org/19745

Yousuf 'Jay' Philips