[PATCH] po/hi.po: Add Hindi Translation

[PATCH] po/hi.po: Add Hindi Translation

[Ramkumar Ramachandra]

Translate a few Git messages to Standard Hindi. It uses a phonetic
script called Devanagari. Focus especially on transforming grammar and
tense in a sane manner to serve as a guideline for future translators.

Note that Devanagari rendering is broken on many applications by
default at this time- the latest Emacs, Vim and Chrome render it
incorrectly.

Signed-off-by: Ramkumar Ramachandra <artagnon@gmail.com>
Cc: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
---
 Note that I don't know Hindi well myself- I read the grammar rules,
 looked up some dictionaries, and used peer review to focus on few
 high-quality translations. The grammar/ tense transformation from
 English is especially non-trivial. I'd encourage an expert Hindi
 speaker to contribute to this.

 To fix rendering in Emacs, see
 http://www.emacswiki.org/emacs/SupportHindi

 po/hi.po |   67 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
 1 files changed, 67 insertions(+), 0 deletions(-)
 create mode 100644 po/hi.po

diff --git a/po/hi.po b/po/hi.po
new file mode 100644
index 0000000..d31e60b
--- /dev/null
+++ b/po/hi.po
@@ -0,0 +1,67 @@
+msgid ""
+msgstr ""
+"Project-Id-Version: Git\n"
+"Report-Msgid-Bugs-To: Git Mailing List <git@vger.kernel.org>\n"
+"Language-Team: Git Mailing List <git@vger.kernel.org>\n"
+"Last-Translator: Ramkumar Ramachandra <artagnon@gmail.com>\n"
+"POT-Creation-Date: 2010-09-05 16:05+0530\n"
+"Language: hi\n"
+"MIME-Version: 1.0\n"
+"Content-Type: text/plain; charset=UTF-8\n"
+"Content-Transfer-Encoding: 8bit\n"
+
+#, c-format
+msgid "Could not make %s writable by group"
+msgstr "%s को समूह द्वारा लिखने योग्य नहीं बना सके"
+
+#, c-format
+msgid "insanely long template name %s"
+msgstr "टेम्पलेट का नाम बहुत लंबा है: %s"
+
+#, c-format
+msgid "cannot stat '%s'"
+msgstr "'%s' को [stat] नहीं कर सके"
+
+#, c-format
+msgid "cannot stat template '%s'"
+msgstr "टेम्पलेट '%s' को [stat] नहीं कर सके"
+
+#, c-format
+msgid "cannot opendir '%s'"
+msgstr "निर्देशिका नहीं खोल सके: '%s'"
+
+#, c-format
+msgid "cannot readlink '%s'"
+msgstr "लिंक नहीं पढ़ सके: '%s'"
+
+#, c-format
+msgid "insanely long symlink %s"
+msgstr "प्रतीकात्मक लिंक बहुत लंबा है: %s"
+
+#, c-format
+msgid "cannot symlink '%s' '%s'"
+msgstr "'%s' से '%s' प्रतीकात्मक लिंक नहीं बना सके"
+
+#, c-format
+msgid "cannot copy '%s' to '%s'"
+msgstr "'%s' का नकल '%s' नहीं बना सके"
+
+#, c-format
+msgid "ignoring template %s"
+msgstr "टेम्पलेट को उपेक्षा कर रहे है: %s"
+
+#, c-format
+msgid "insanely long template path %s"
+msgstr "टेम्पलेट का आकार बहुत लंबा है: %s"
+
+#, c-format
+msgid "templates not found %s"
+msgstr "टेम्पलेट नहीं मिला: %s"
+
+#, c-format
+msgid "not copying templates of a wrong format version %d from '%s'"
+msgstr "'%s' से गलत स्वरूप संस्करण %d का टेम्पलेट नकल नहीं कर रहे है"
+
+#, c-format
+msgid "insane git directory %s"
+msgstr "खराब गिट निर्देशिका %s"
-- 
1.7.2.2.409.gdbb11.dirty

Re: [PATCH] po/hi.po: Add Hindi Translation

[Ævar Arnfjörð Bjarmason]

On Sun, Sep 5, 2010 at 18:27, Ramkumar Ramachandra <artagnon@gmail.com> wrote:
> Translate a few Git messages to Standard Hindi. It uses a phonetic
> script called Devanagari. Focus especially on transforming grammar and
> tense in a sane manner to serve as a guideline for future translators.
>
> Note that Devanagari rendering is broken on many applications by
> default at this time- the latest Emacs, Vim and Chrome render it
> incorrectly.
>
> Signed-off-by: Ramkumar Ramachandra <artagnon@gmail.com>
> Cc: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
> ---
>  Note that I don't know Hindi well myself- I read the grammar rules,
>  looked up some dictionaries, and used peer review to focus on few
>  high-quality translations. The grammar/ tense transformation from
>  English is especially non-trivial. I'd encourage an expert Hindi
>  speaker to contribute to this.

I don't know Hindi either, but the gettext headers and everything else
about this looks OK. Thanks for contributing to translation.

Acked-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>

Re: [PATCH] po/hi.po: Add Hindi Translation

[Sitaram Chamarty]

On Sun, Sep 5, 2010 at 11:57 PM, Ramkumar Ramachandra
<artagnon@gmail.com> wrote:

>  Note that I don't know Hindi well myself- I read the grammar rules,
>  looked up some dictionaries, and used peer review to focus on few
>  high-quality translations. The grammar/ tense transformation from
>  English is especially non-trivial. I'd encourage an expert Hindi
>  speaker to contribute to this.

looks OK to me.  Some of those words will never be encountered in
colloquial Hindi (I had to think for a bit on निर्देशिका and
प्रतीकात्मक लिंक!) , but that's the way it always is.  We end up just
using the English word within Hindi/Telugu/whatever :-)

Re: [PATCH] po/hi.po: Add Hindi Translation

[Ramkumar Ramachandra]

Hi Sitaram,

Sitaram Chamarty writes:
> looks OK to me.  Some of those words will never be encountered in
> colloquial Hindi (I had to think for a bit on निर्देशिका and
> प्रतीकात्मक लिंक!) , but that's the way it always is.  We end up just
> using the English word within Hindi/Telugu/whatever :-)

Thanks for the feedback. Yes, that's true- even among my friends, only
a few expert Hindi speakers were able to follow hindi.po
completely. However, I figured that only those people who are more
comfortable in Hindi than English will actually use hindi.po, so we
might as well do it right :)

I also discovered another thing- even expert Hindi speakers don't seem
to understand Hindi grammar very well :p After a few hours of
studying, I was able to correct a few more mistakes that others
weren't able to spot. Here's the diff.

Junio: Please squash this into the first revision of the patch.

diff --git a/po/hi.po b/po/hi.po
index d31e60b..a5464b3 100644
--- a/po/hi.po
+++ b/po/hi.po
@@ -12,11 +12,11 @@ msgstr ""
 
 #, c-format
 msgid "Could not make %s writable by group"
-msgstr "%s को समूह द्वारा लिखने योग्य नहीं बना सके"
+msgstr "%s को समूह द्वारा लिखने की इजाज़त नही दे सके"
 
 #, c-format
 msgid "insanely long template name %s"
-msgstr "टेम्पलेट का नाम बहुत लंबा है: %s"
+msgstr "टेम्पलेट %s का नाम बहुत लंबा है"
 
 #, c-format
 msgid "cannot stat '%s'"
@@ -28,40 +28,40 @@ msgstr "टेम्पलेट '%s' को [stat] नहीं कर सक
 
 #, c-format
 msgid "cannot opendir '%s'"
-msgstr "निर्देशिका नहीं खोल सके: '%s'"
+msgstr "निर्देशिका '%s' नहीं खोल सके"
 
 #, c-format
 msgid "cannot readlink '%s'"
-msgstr "लिंक नहीं पढ़ सके: '%s'"
+msgstr "लिंक '%s' नहीं पढ़ सके"
 
 #, c-format
 msgid "insanely long symlink %s"
-msgstr "प्रतीकात्मक लिंक बहुत लंबा है: %s"
+msgstr "प्रतीकात्मक लिंक %s बहुत लंबा है"
 
 #, c-format
 msgid "cannot symlink '%s' '%s'"
-msgstr "'%s' से '%s' प्रतीकात्मक लिंक नहीं बना सके"
+msgstr "'%s' का प्रतीकात्मक लिंक '%s' में नहीं बना सके"
 
 #, c-format
 msgid "cannot copy '%s' to '%s'"
-msgstr "'%s' का नकल '%s' नहीं बना सके"
+msgstr "'%s' का अनुकृति '%s' में नहीं कर सके"
 
 #, c-format
 msgid "ignoring template %s"
-msgstr "टेम्पलेट को उपेक्षा कर रहे है: %s"
+msgstr "टेम्पलेट %s का उपेक्षा कर रहे है"
 
 #, c-format
 msgid "insanely long template path %s"
-msgstr "टेम्पलेट का आकार बहुत लंबा है: %s"
+msgstr "टेम्पलेट %s का आकार बहुत लंबा है"
 
 #, c-format
 msgid "templates not found %s"
-msgstr "टेम्पलेट नहीं मिला: %s"
+msgstr "टेम्पलेट %s नहीं मिला"
 
 #, c-format
 msgid "not copying templates of a wrong format version %d from '%s'"
-msgstr "'%s' से गलत स्वरूप संस्करण %d का टेम्पलेट नकल नहीं कर रहे है"
+msgstr "'%s' से गलत स्वरूप संस्करण %d के टेम्पलेट्स के नकल नहीं कर रहे है"
 
 #, c-format
 msgid "insane git directory %s"
-msgstr "खराब गिट निर्देशिका %s"
+msgstr "%s खराब गिट निर्देशिका है"

Re: [PATCH] po/hi.po: Add Hindi Translation

[Antriksh Pany]

Hi Ram

Nice to see some Hindi around here ;P

A few comments inline.

On Mon, Sep 6, 2010 at 9:59 AM, Ramkumar Ramachandra <artagnon@gmail.com> wrote:
> Hi Sitaram,
>
> Sitaram Chamarty writes:
>> looks OK to me.  Some of those words will never be encountered in
>> colloquial Hindi (I had to think for a bit on निर्देशिका and
>> प्रतीकात्मक लिंक!) , but that's the way it always is.  We end up just
>> using the English word within Hindi/Telugu/whatever :-)
>
> Thanks for the feedback. Yes, that's true- even among my friends, only
> a few expert Hindi speakers were able to follow hindi.po
> completely. However, I figured that only those people who are more
> comfortable in Hindi than English will actually use hindi.po, so we
> might as well do it right :)
>
> I also discovered another thing- even expert Hindi speakers don't seem
> to understand Hindi grammar very well :p After a few hours of
> studying, I was able to correct a few more mistakes that others
> weren't able to spot. Here's the diff.
>
> Junio: Please squash this into the first revision of the patch.
>
> diff --git a/po/hi.po b/po/hi.po
> index d31e60b..a5464b3 100644
> --- a/po/hi.po
> +++ b/po/hi.po
> @@ -12,11 +12,11 @@ msgstr ""
>
>  #, c-format
>  msgid "Could not make %s writable by group"
> -msgstr "%s को समूह द्वारा लिखने योग्य नहीं बना सके"
> +msgstr "%s को समूह द्वारा लिखने की इजाज़त नही दे सके"
>
>  #, c-format
>  msgid "insanely long template name %s"
> -msgstr "टेम्पलेट का नाम बहुत लंबा है: %s"
> +msgstr "टेम्पलेट %s का नाम बहुत लंबा है"
bahut -> atyant?

>
>  #, c-format
>  msgid "cannot stat '%s'"
> @@ -28,40 +28,40 @@ msgstr "टेम्पलेट '%s' को [stat] नहीं कर सक
>
>  #, c-format
>  msgid "cannot opendir '%s'"
> -msgstr "निर्देशिका नहीं खोल सके: '%s'"
> +msgstr "निर्देशिका '%s' नहीं खोल सके"
>
>  #, c-format
>  msgid "cannot readlink '%s'"
> -msgstr "लिंक नहीं पढ़ सके: '%s'"
> +msgstr "लिंक '%s' नहीं पढ़ सके"
>
>  #, c-format
>  msgid "insanely long symlink %s"
> -msgstr "प्रतीकात्मक लिंक बहुत लंबा है: %s"
> +msgstr "प्रतीकात्मक लिंक %s बहुत लंबा है"
Same suggestion: bahut -> atyant?
Also, dropping the 'है' might be crisper and closer to the English version.

>
>  #, c-format
>  msgid "cannot symlink '%s' '%s'"
> -msgstr "'%s' से '%s' प्रतीकात्मक लिंक नहीं बना सके"
> +msgstr "'%s' का प्रतीकात्मक लिंक '%s' में नहीं बना सके"
>
>  #, c-format
>  msgid "cannot copy '%s' to '%s'"
> -msgstr "'%s' का नकल '%s' नहीं बना सके"
> +msgstr "'%s' का अनुकृति '%s' में नहीं कर सके"
ka -> ki

>
>  #, c-format
>  msgid "ignoring template %s"
> -msgstr "टेम्पलेट को उपेक्षा कर रहे है: %s"
> +msgstr "टेम्पलेट %s का उपेक्षा कर रहे है"
ka -> ki

>
>  #, c-format
>  msgid "insanely long template path %s"
> -msgstr "टेम्पलेट का आकार बहुत लंबा है: %s"
> +msgstr "टेम्पलेट %s का आकार बहुत लंबा है"
Again: bahut -> atyant. And again, 'है' could be dropped perhaps.

>
>  #, c-format
>  msgid "templates not found %s"
> -msgstr "टेम्पलेट नहीं मिला: %s"
> +msgstr "टेम्पलेट %s नहीं मिला"
>
>  #, c-format
>  msgid "not copying templates of a wrong format version %d from '%s'"
> -msgstr "'%s' से गलत स्वरूप संस्करण %d का टेम्पलेट नकल नहीं कर रहे है"
> +msgstr "'%s' से गलत स्वरूप संस्करण %d के टेम्पलेट्स के नकल नहीं कर रहे है"
Had changed नकल to अनुकृति earlier. Not doing that here?

>
>  #, c-format
>  msgid "insane git directory %s"
> -msgstr "खराब गिट निर्देशिका %s"
> +msgstr "%s खराब गिट निर्देशिका है"
I think the first one was better.


Having gone through this, I find it hard to think that anyone is
really going to use the Hindi translation. Any Indian who would be
involved with git would almost certainly be involved with it in the
English language, particularly given that some of the Hindi words are
fairly hard to follow. Also, when it comes to the Sciences (and in
particular, Computers) Indians (at least the Hindi speaking
population) have directly adopted the English versions instead of
seeking appropriate Hindi ones.

If we are going to have a hindi.po, I think it may be altogether
better to drop the unusual Hindi words, and literally use the English
word written in Hindi.. the way you did for words like 'templates',
'link' etc. Other candidates could be 'directory', 'format',
'version', 'copy', 'symbolic (link)', 'group', 'read', 'write'. I know
that translation for the last few words are not complicated. However,
the subtle meanings of these words in the context of computers gets
lost when translated in Hindi. (For example, 'read' and 'write', while
being similar to real-life usage, have in fact taken very specific
meanings in the context of computer science.)

Cheers
Antriksh

Re: [PATCH] po/hi.po: Add Hindi Translation

[Ramkumar Ramachandra]

Hi Antriksh,

Antriksh Pany writes:
> Nice to see some Hindi around here ;P
> 
> A few comments inline.

Thanks for the wonderful feedback! :)

> >  #, c-format
> >  msgid "insanely long template name %s"
> > -msgstr "टेम्पलेट का नाम बहुत लंबा है: %s"
> > +msgstr "टेम्पलेट %s का नाम बहुत लंबा है"
> bahut -> atyant?

Yes, definitely. बहुत did feel out of place here.

> >  #, c-format
> >  msgid "insanely long symlink %s"
> > -msgstr "प्रतीकात्मक लिंक बहुत लंबा है: %s"
> > +msgstr "प्रतीकात्मक लिंक %s बहुत लंबा है"
> Same suggestion: bahut -> atyant?
> Also, dropping the 'है' might be crisper and closer to the English version.

Okay.

> >  #, c-format
> >  msgid "cannot copy '%s' to '%s'"
> > -msgstr "'%s' का नकल '%s' नहीं बना सके"
> > +msgstr "'%s' का अनुकृति '%s' में नहीं कर सके"
> ka -> ki

Ah, yes. Gender ambiguity in Hindi. I also felt it should be की. Can
you explain why though? '%s' a noun here; have you chosen it to be
feminine? But अनुकृति is a noun as well, and quite obviously
feminine. Is that why?

> >  #, c-format
> >  msgid "ignoring template %s"
> > -msgstr "टेम्पलेट को उपेक्षा कर रहे है: %s"
> > +msgstr "टेम्पलेट %s का उपेक्षा कर रहे है"
> ka -> ki

Similar confusion- I'd like an explanation if possible. उपेक्षा is a
verb, and I'm not sure how it transforms when gender is afflicted on
it. I have chosen टेम्पलेट to be masculine- do you think it should be
feminine?

> >  #, c-format
> >  msgid "insanely long template path %s"
> > -msgstr "टेम्पलेट का आकार बहुत लंबा है: %s"
> > +msgstr "टेम्पलेट %s का आकार बहुत लंबा है"
> Again: bahut -> atyant. And again, 'है' could be dropped perhaps.

Ok.

> >  #, c-format
> >  msgid "not copying templates of a wrong format version %d from '%s'"
> > -msgstr "'%s' से गलत स्वरूप संस्करण %d का टेम्पलेट नकल नहीं कर रहे है"
> > +msgstr "'%s' से गलत स्वरूप संस्करण %d के टेम्पलेट्स के नकल नहीं कर रहे है"
> Had changed नकल to अनुकृति earlier. Not doing that here?

While नकल is definitely the easier word to use, it's a verb. अनुकृति is
a noun on the other hand- I had to introduce it into the previous
sentence because "Make a नकल of X in Y" would be technically
incorrect.

> >  #, c-format
> >  msgid "insane git directory %s"
> > -msgstr "खराब गिट निर्देशिका %s"
> > +msgstr "%s खराब गिट निर्देशिका है"
> I think the first one was better.

Ok.

> Having gone through this, I find it hard to think that anyone is
> really going to use the Hindi translation. Any Indian who would be
> involved with git would almost certainly be involved with it in the
> English language, particularly given that some of the Hindi words are
> fairly hard to follow. Also, when it comes to the Sciences (and in
> particular, Computers) Indians (at least the Hindi speaking
> population) have directly adopted the English versions instead of
> seeking appropriate Hindi ones.

I understand.

> If we are going to have a hindi.po, I think it may be altogether
> better to drop the unusual Hindi words, and literally use the English
> word written in Hindi.. the way you did for words like 'templates',
> 'link' etc. Other candidates could be 'directory', 'format',
> 'version', 'copy', 'symbolic (link)', 'group', 'read', 'write'. I know
> that translation for the last few words are not complicated. However,
> the subtle meanings of these words in the context of computers gets
> lost when translated in Hindi. (For example, 'read' and 'write', while
> being similar to real-life usage, have in fact taken very specific
> meanings in the context of computer science.)

A number of my friends pointed out the same thing. I don't completely
agree though. In my opinion, the translations are really just fun, and
give the user a warm feeling- it's quite unlikely that a person
(atleast Indian) who doesn't understand English will use Git in the
first place.

Personally, I find English words written in Devanagiri script quite
annoying: there is really no elegance or point in translating "Could
not stat copy of symbolic link" to "सिम्बोलिक लिंक का कॉपी स्टाट नहीं कर
सके". A person who uses hi.po is doing it mainly for recreational
purposes, and we'd like to give that person a beautiful and elegant
experience, instead of dumbly writing all the nouns in Devanagiri
script and inserting a few trivial filler verbs.

While I probably took this too far with words like निर्देशिका, I'd like
to be able to find some kind of compromise- I'll submit another
revision shortly using a combination of real Hindi words, English
words written in Devanagiri script, and English words expressed
plainly (like "[stat]").

I wasn't joking when I said that I don't know Hindi. I've never used
applications that speak anything but English, and have never been
involved with translation before. Thanks for poking- I'll email some
Indians involved with translation projects hi.po and ask them for
feedback.

Thanks.

-- Ram

Re: [PATCH] po/hi.po: Add Hindi Translation

[Antriksh Pany]

Hi Ram

Some more comments (spotted a few more possible improvements) and
responses inline.

On Tue, Sep 7, 2010 at 1:25 PM, Ramkumar Ramachandra <artagnon@gmail.com> wrote:
> Hi Antriksh,
>
> Antriksh Pany writes:
>> Nice to see some Hindi around here ;P
>>
>> A few comments inline.
>
> Thanks for the wonderful feedback! :)
>
>> >  #, c-format
>> >  msgid "insanely long template name %s"
>> > -msgstr "टेम्पलेट का नाम बहुत लंबा है: %s"
>> > +msgstr "टेम्पलेट %s का नाम बहुत लंबा है"
>> bahut -> atyant?
>
> Yes, definitely. बहुत did feel out of place here.
>
>> >  #, c-format
>> >  msgid "insanely long symlink %s"
>> > -msgstr "प्रतीकात्मक लिंक बहुत लंबा है: %s"
>> > +msgstr "प्रतीकात्मक लिंक %s बहुत लंबा है"
>> Same suggestion: bahut -> atyant?
>> Also, dropping the 'है' might be crisper and closer to the English
>> version.
>
> Okay.
>
>> >  #, c-format
>> >  msgid "cannot copy '%s' to '%s'"
>> > -msgstr "'%s' का नकल '%s' नहीं बना सके"
>> > +msgstr "'%s' का अनुकृति '%s' में नहीं कर सके"
>> ka -> ki
>
> Ah, yes. Gender ambiguity in Hindi. I also felt it should be की. Can
> you explain why though? '%s' a noun here; have you chosen it to be
> feminine? But अनुकृति is a noun as well, and quite obviously
> feminine. Is that why?

Yes, अनुकृति being feminine is what causes 'ki' to be used here.

>
>> >  #, c-format
>> >  msgid "ignoring template %s"
>> > -msgstr "टेम्पलेट को उपेक्षा कर रहे है: %s"
>> > +msgstr "टेम्पलेट %s का उपेक्षा कर रहे है"
>> ka -> ki
>
> Similar confusion- I'd like an explanation if possible. उपेक्षा is a
> verb, and I'm not sure how it transforms when gender is afflicted on
> it. I have chosen टेम्पलेट to be masculine- do you think it should be
> feminine?

Again, उपेक्षा is determining the gender usage.
Also, not sure if this is a good word to use. उपेक्षा typically has a negative
connotation, so does not seem to fit well here. अनदेखा (while not being
perfect either) may be better?

>
>> >  #, c-format
>> >  msgid "insanely long template path %s"
>> > -msgstr "टेम्पलेट का आकार बहुत लंबा है: %s"
>> > +msgstr "टेम्पलेट %s का आकार बहुत लंबा है"
>> Again: bahut -> atyant. And again, 'है' could be dropped perhaps.
>
> Ok.

On second thoughts, आकार is definitely not close enough to 'path'.
Can we think of something else?

>
>> >  #, c-format
>> >  msgid "not copying templates of a wrong format version %d from '%s'"
>> > -msgstr "'%s' से गलत स्वरूप संस्करण %d का टेम्पलेट नकल नहीं कर रहे है"
>> > +msgstr "'%s' से गलत स्वरूप संस्करण %d के टेम्पलेट्स के नकल नहीं कर रहे
>> > है"
>> Had changed नकल to अनुकृति earlier. Not doing that here?
>
> While नकल is definitely the easier word to use, it's a verb. अनुकृति is
> a noun on the other hand- I had to introduce it into the previous
> sentence because "Make a नकल of X in Y" would be technically
> incorrect.
>

'टेम्पलेट्स के ' -> 'ki' would have to be used here. 'नकल' is determining
gender here :)

>> >  #, c-format
>> >  msgid "insane git directory %s"
>> > -msgstr "खराब गिट निर्देशिका %s"
>> > +msgstr "%s खराब गिट निर्देशिका है"
>> I think the first one was better.
>
> Ok.
>
>> Having gone through this, I find it hard to think that anyone is
>> really going to use the Hindi translation. Any Indian who would be
>> involved with git would almost certainly be involved with it in the
>> English language, particularly given that some of the Hindi words are
>> fairly hard to follow. Also, when it comes to the Sciences (and in
>> particular, Computers) Indians (at least the Hindi speaking
>> population) have directly adopted the English versions instead of
>> seeking appropriate Hindi ones.
>
> I understand.
>
>> If we are going to have a hindi.po, I think it may be altogether
>> better to drop the unusual Hindi words, and literally use the English
>> word written in Hindi.. the way you did for words like 'templates',
>> 'link' etc. Other candidates could be 'directory', 'format',
>> 'version', 'copy', 'symbolic (link)', 'group', 'read', 'write'. I know
>> that translation for the last few words are not complicated. However,
>> the subtle meanings of these words in the context of computers gets
>> lost when translated in Hindi. (For example, 'read' and 'write', while
>> being similar to real-life usage, have in fact taken very specific
>> meanings in the context of computer science.)
>
> A number of my friends pointed out the same thing. I don't completely
> agree though. In my opinion, the translations are really just fun, and
> give the user a warm feeling- it's quite unlikely that a person
> (atleast Indian) who doesn't understand English will use Git in the
> first place.
>
> Personally, I find English words written in Devanagiri script quite
> annoying: there is really no elegance or point in translating "Could
> not stat copy of symbolic link" to "सिम्बोलिक लिंक का कॉपी स्टाट नहीं कर
> सके". A person who uses hi.po is doing it mainly for recreational
> purposes, and we'd like to give that person a beautiful and elegant
> experience, instead of dumbly writing all the nouns in Devanagiri
> script and inserting a few trivial filler verbs.
>

Well, I really thought that the purpose of translation is more utilitarian
than anything else. And I would still count myself among those friends
of yours who either don't see many Indians using this, or see some value
only if words which don't have a direct reference in Hindi are transliterated
:)

There are other languages where effort is taken to find equivalents to modern
terms, and hence don't need a lot of transliterations. That, unfortunately, is
not the case with Hindi :(

> While I probably took this too far with words like निर्देशिका, I'd like
> to be able to find some kind of compromise- I'll submit another
> revision shortly using a combination of real Hindi words, English
> words written in Devanagiri script, and English words expressed
> plainly (like "[stat]").
>
> I wasn't joking when I said that I don't know Hindi. I've never used
> applications that speak anything but English, and have never been
> involved with translation before. Thanks for poking- I'll email some
> Indians involved with translation projects hi.po and ask them for
> feedback.
>

Really appreciate the effort you have put in on this, despite not being
a native Hindi speaker. Sorry for so many nits :)

Cheers
Antriksh

Re: [PATCH] po/hi.po: Add Hindi Translation

[Ramkumar Ramachandra]

Hi Antriksh,

Antriksh Pany writes:
> Some more comments (spotted a few more possible improvements) and
> responses inline.

Thanks for additionally clearing my doubts :)

> >> >  #, c-format
> >> >  msgid "cannot copy '%s' to '%s'"
> >> > -msgstr "'%s' का नकल '%s' नहीं बना सके"
> >> > +msgstr "'%s' का अनुकृति '%s' में नहीं कर सके"
> >> ka -> ki
> >
> > Ah, yes. Gender ambiguity in Hindi. I also felt it should be की. Can
> > you explain why though? '%s' a noun here; have you chosen it to be
> > feminine? But अनुकृति is a noun as well, and quite obviously
> > feminine. Is that why?
> 
> Yes, अनुकृति being feminine is what causes 'ki' to be used here.

Got it.

> >> >  #, c-format
> >> >  msgid "ignoring template %s"
> >> > -msgstr "टेम्पलेट को उपेक्षा कर रहे है: %s"
> >> > +msgstr "टेम्पलेट %s का उपेक्षा कर रहे है"
> >> ka -> ki
> >
> > Similar confusion- I'd like an explanation if possible. उपेक्षा is a
> > verb, and I'm not sure how it transforms when gender is afflicted on
> > it. I have chosen टेम्पलेट to be masculine- do you think it should be
> > feminine?
> 
> Again, उपेक्षा is determining the gender usage.
> Also, not sure if this is a good word to use. उपेक्षा typically has a negative
> connotation, so does not seem to fit well here. अनदेखा (while not being
> perfect either) may be better?

Actually, this was my intent. I considered अनदेखा in the beginning, but
I thought उपेक्षा was closer. अनदेखा sounds like "didn't look", but I was
aiming for a more aggressive word for conciously ignoring.

> >> >  #, c-format
> >> >  msgid "insanely long template path %s"
> >> > -msgstr "टेम्पलेट का आकार बहुत लंबा है: %s"
> >> > +msgstr "टेम्पलेट %s का आकार बहुत लंबा है"
> >> Again: bahut -> atyant. And again, 'है' could be dropped perhaps.
> >
> > Ok.
> 
> On second thoughts, आकार is definitely not close enough to 'path'.
> Can we think of something else?

A literal translation like पथ is probably not the way to go. राह is
only marginally better. Although आकार literally means "form", I also
think it's not close enough. If we run out of ideas, we have to
transliterate.

> >> >  #, c-format
> >> >  msgid "not copying templates of a wrong format version %d from '%s'"
> >> > -msgstr "'%s' से गलत स्वरूप संस्करण %d का टेम्पलेट नकल नहीं कर रहे है"
> >> > +msgstr "'%s' से गलत स्वरूप संस्करण %d के टेम्पलेट्स के नकल नहीं कर रहे
> >> > है"
> >> Had changed नकल to अनुकृति earlier. Not doing that here?
> >
> > While नकल is definitely the easier word to use, it's a verb. अनुकृति is
> > a noun on the other hand- I had to introduce it into the previous
> > sentence because "Make a नकल of X in Y" would be technically
> > incorrect.
> >
> 
> 'टेम्पलेट्स के ' -> 'ki' would have to be used here. 'नकल' is determining
> gender here :)

Nice catch :)

> Well, I really thought that the purpose of translation is more utilitarian
> than anything else. And I would still count myself among those friends
> of yours who either don't see many Indians using this, or see some value
> only if words which don't have a direct reference in Hindi are transliterated
> :)
> 
> There are other languages where effort is taken to find equivalents to modern
> terms, and hence don't need a lot of transliterations. That, unfortunately, is
> not the case with Hindi :(

Yeah, translation definitely hasn't caught up here- especially with
everyone aggressively pursuing English education.

> Really appreciate the effort you have put in on this, despite not being
> a native Hindi speaker. Sorry for so many nits :)

Thanks for the nits. Being a grammar nazi can't be easy :)
I'll submit another revision with some amendments soon.

-- Ram

Re: [PATCH] po/hi.po: Add Hindi Translation

[Ramkumar Ramachandra]

Hi Thiago,

[+cc: list, because I think others will find it useful as well]

Thiago Farina writes:
> How did you do this translation? I would like to do the Portuguese
> Brazilian translation. I have tried the poedit, but I couldn't make it
> work yet.

Here's a quick tutorial to get started:
1. Get Ævar's fork of git.git from [1] and checkout the `ab/i18n-all`
   branch.
2. Read the instructions in po/README to create the actual .po file.
3. Get an editor that has transliteration support. For example, for
   Emacs and Hindi, execute (set-input-method devanagari-itrans) and
   use C-\ to switch between the input methods. 
4. If you're not able to find a suitable editor or otherwise, use
   Google Transliterate [2].
5. Optionally, copy out some additional headers from `po/is.po` etc.
6. Translate and send the patch to the list.

Hope that helps.

-- Ram

[1]: git://github.com/avar/git.git
[2]: http://google.com/transliterate