Opinions on the localization of the help pages

classic Classic list List threaded Threaded
8 messages Options
sylvestre sylvestre
Reply | Threaded
Open this post in threaded view
|

Opinions on the localization of the help pages

Hello guys,

While Scilab itself is pretty well translated, it is not really the case
of it documentation.
In English, the reference, we have 2393 xml files (usually, more or less
a page). It is docbook files.
We have translations in French, portugues do brazil & japanese.

Clement suggested to use xml2po & po2xml to simplify the work:
http://bugzilla.scilab.org/show_bug.cgi?id=9481
and pluggin it with launchpad.

What is your opinions on this ?

Thanks
S


Yuri Chornoivan Yuri Chornoivan
Reply | Threaded
Open this post in threaded view
|

Re: Opinions on the localization of the help pages

написане Sat, 14 May 2011 17:08:45 +0300, sylvestre [via Scilab / Xcos -  
Mailing Lists Archives] <[hidden email]>:

>
>
> Hello guys,
>
> While Scilab itself is pretty well translated, it is not really the case
> of it documentation.
> In English, the reference, we have 2393 xml files (usually, more or less
> a page). It is docbook files.
> We have translations in French, portugues do brazil & japanese.
>
> Clement suggested to use xml2po & po2xml to simplify the work:
> http://bugzilla.scilab.org/show_bug.cgi?id=9481
> and pluggin it with launchpad.
>
> What is your opinions on this ?

Hi!

It seems that it will be problematic to make the full translation during  
the next release cycle.

On the other hand po2xml generates docbook only for 100% translated PO  
that is 100% XML-compliant.

It is good for consistency, but Rosetta do not check the XML (it is well  
known that Ubuntu docs are broken for many languages because the  
translators ignore the guidelines or use the wrong guidelines). In KDE we  
use pology+Lokalize to debug POs with XML code, but it can be the overkill  
for Scilab.

Thus, The following questions have to be answered:

1. The improper POs can be excluded from docbook generation, but how to  
send the message with the errors to the translators?

2. To generate docbooks from the partial translation we can modify po2xml  
(patches can be found in kde-i18n-doc, I can give the links if needed) or  
use po4a (it seems to me that in this case it will be also better to use  
po4a for generate POTs generation). So what the proper level of the  
translation to include it into the release?

3. If there is no translation English docbook should be shown. Is this  
feasible?

If these issues can be solved, I think there are no other obstacles to  
make the docs translatable.

Thanks in advance for the answers.

Best regards,
Yuri
Kenneth Nielsen Kenneth Nielsen
Reply | Threaded
Open this post in threaded view
|

Re: Opinions on the localization of the help pages

Hey

> Hi!
>
> It seems that it will be problematic to make the full translation during
> the next release cycle.
>
> On the other hand po2xml generates docbook only for 100% translated PO
> that is 100% XML-compliant.
> It is good for consistency, but Rosetta do not check the XML (it is well
> known that Ubuntu docs are broken for many languages because the
> translators ignore the guidelines or use the wrong guidelines). In KDE we
> use pology+Lokalize to debug POs with XML code, but it can be the overkill
> for Scilab.

Alternatively you can also use gtxml (GetTextXML) from the pyg3t
package (of which I'm a co-author, this is really just shameless
adverticement ;)). gtxml is written by a fellow Danish translator Ask
H. Larsen. It is a commandline tool that generates helpful error
messages, so it can be used e.g. for checking the xml of all the
localised files and sending out a combined email with errors for all
of them.

> Thus, The following questions have to be answered:
>
> 1. The improper POs can be excluded from docbook generation, but how to
> send the message with the errors to the translators?

gtxml could produce the errors, the communication is a bit more
tricky. It might also be hooked in a build script to exclude broken
files at build time, but I guess maybe the proper way to fix the
problem, is to do the check as a commit hook, so the bad files never
make it in the source tree.

Regards Kenneth

> 2. To generate docbooks from the partial translation we can modify po2xml
> (patches can be found in kde-i18n-doc, I can give the links if needed) or
> use po4a (it seems to me that in this case it will be also better to use
> po4a for generate POTs generation). So what the proper level of the
> translation to include it into the release?
>
> 3. If there is no translation English docbook should be shown. Is this
> feasible?
>
> If these issues can be solved, I think there are no other obstacles to
> make the docs translatable.
>
> Thanks in advance for the answers.
>
> Best regards,
> Yuri
>
> ________________________________
> View this message in context: Re: Opinions on the localization of the help
> pages
> Sent from the Scilab localization - Mailing Lists Archives mailing list
> archive at Nabble.com.
>
Kenneth Nielsen Kenneth Nielsen
Reply | Threaded
Open this post in threaded view
|

Re: Opinions on the localization of the help pages

2011/5/14 Kenneth Nielsen <[hidden email]>:

> Hey
>
>> Hi!
>>
>> It seems that it will be problematic to make the full translation during
>> the next release cycle.
>>
>> On the other hand po2xml generates docbook only for 100% translated PO
>> that is 100% XML-compliant.
>> It is good for consistency, but Rosetta do not check the XML (it is well
>> known that Ubuntu docs are broken for many languages because the
>> translators ignore the guidelines or use the wrong guidelines). In KDE we
>> use pology+Lokalize to debug POs with XML code, but it can be the overkill
>> for Scilab.
>
> Alternatively you can also use gtxml (GetTextXML) from the pyg3t
> package (of which I'm a co-author, this is really just shameless
> adverticement ;)). gtxml is written by a fellow Danish translator Ask
> H. Larsen. It is a commandline tool that generates helpful error
> messages, so it can be used e.g. for checking the xml of all the
> localised files and sending out a combined email with errors for all
> of them.
>
>> Thus, The following questions have to be answered:
>>
>> 1. The improper POs can be excluded from docbook generation, but how to
>> send the message with the errors to the translators?
>
> gtxml could produce the errors, the communication is a bit more
> tricky. It might also be hooked in a build script to exclude broken
> files at build time, but I guess maybe the proper way to fix the

Uhh I forgot to say that this has never been done though. It is a
relatively new tool.

> problem, is to do the check as a commit hook, so the bad files never
> make it in the source tree.
>
> Regards Kenneth
>
>> 2. To generate docbooks from the partial translation we can modify po2xml
>> (patches can be found in kde-i18n-doc, I can give the links if needed) or
>> use po4a (it seems to me that in this case it will be also better to use
>> po4a for generate POTs generation). So what the proper level of the
>> translation to include it into the release?
>>
>> 3. If there is no translation English docbook should be shown. Is this
>> feasible?
>>
>> If these issues can be solved, I think there are no other obstacles to
>> make the docs translatable.
>>
>> Thanks in advance for the answers.
>>
>> Best regards,
>> Yuri
>>
>> ________________________________
>> View this message in context: Re: Opinions on the localization of the help
>> pages
>> Sent from the Scilab localization - Mailing Lists Archives mailing list
>> archive at Nabble.com.
>>
>
sylvestre sylvestre
Reply | Threaded
Open this post in threaded view
|

Re: Opinions on the localization of the help pages

In reply to this post by Yuri Chornoivan
Le samedi 14 mai 2011 à 10:02 -0700, Yuri Chornoivan a écrit :

> написане Sat, 14 May 2011 17:08:45 +0300, sylvestre [via Scilab / Xcos
> -  
> Mailing Lists Archives] <[hidden email]>:
>
> >
> >
> > Hello guys,
> >
> > While Scilab itself is pretty well translated, it is not really the
> case
> > of it documentation.
> > In English, the reference, we have 2393 xml files (usually, more or
> less
> > a page). It is docbook files.
> > We have translations in French, portugues do brazil & japanese.
> >
> > Clement suggested to use xml2po & po2xml to simplify the work:
> > http://bugzilla.scilab.org/show_bug.cgi?id=9481
> > and pluggin it with launchpad.
> >
> > What is your opinions on this ?
>
> Hi!
>
> It seems that it will be problematic to make the full translation
> during  
> the next release cycle.
It is not that I am expecting.
I am more thinking about good tool to manage that.


> Thus, The following questions have to be answered:
>
> 1. The improper POs can be excluded from docbook generation, but how
> to  
> send the message with the errors to the translators?
It is already a problem with the current localization process.


> 2. To generate docbooks from the partial translation we can modify
> po2xml  
> (patches can be found in kde-i18n-doc, I can give the links if needed)
> or  
> use po4a (it seems to me that in this case it will be also better to
> use  
> po4a for generate POTs generation). So what the proper level of the  
> translation to include it into the release?
At least a third of the documentation translated would be nice. What is
your opinion on this ?

> 3. If there is no translation English docbook should be shown. Is
> this  
> feasible?
The english original doc will be available. For example:
#: mlist.xml:50(para)
msgid "The semantic of the extraction and insertion syntax should be
given by an <link linkend=\"overloading\">overloading</link> functions."
msgstr ""


Sylvestre


sylvestre sylvestre
Reply | Threaded
Open this post in threaded view
|

Re: Opinions on the localization of the help pages

In reply to this post by Kenneth Nielsen
Le samedi 14 mai 2011 à 23:13 +0200, Kenneth Nielsen a écrit :

> Hey
>
> > Hi!
> >
> > It seems that it will be problematic to make the full translation during
> > the next release cycle.
> >
> > On the other hand po2xml generates docbook only for 100% translated PO
> > that is 100% XML-compliant.
> > It is good for consistency, but Rosetta do not check the XML (it is well
> > known that Ubuntu docs are broken for many languages because the
> > translators ignore the guidelines or use the wrong guidelines). In KDE we
> > use pology+Lokalize to debug POs with XML code, but it can be the overkill
> > for Scilab.
>
> Alternatively you can also use gtxml (GetTextXML) from the pyg3t
> package (of which I'm a co-author, this is really just shameless
> adverticement ;)). gtxml is written by a fellow Danish translator Ask
> H. Larsen. It is a commandline tool that generates helpful error
> messages, so it can be used e.g. for checking the xml of all the
> localised files and sending out a combined email with errors for all
> of them.
Do you have some documents and use cases of it ?

thanks,
S


Yuri Chornoivan Yuri Chornoivan
Reply | Threaded
Open this post in threaded view
|

Re: Opinions on the localization of the help pages

In reply to this post by sylvestre
написане Wed, 18 May 2011 18:42:29 +0300, sylvestre [via Scilab / Xcos -  
Mailing Lists Archives] <[hidden email]>:

>> 2. To generate docbooks from the partial translation we can modify
>> po2xml
>> (patches can be found in kde-i18n-doc, I can give the links if needed)
>> or
>> use po4a (it seems to me that in this case it will be also better to
>> use
>> po4a for POTs generation). So what is the proper level of the
>> translation to include it into the release?
> At least a third of the documentation translated would be nice. What is
> your opinion on this ?

That is good for me. :)
Kenneth Nielsen Kenneth Nielsen
Reply | Threaded
Open this post in threaded view
|

Re: Opinions on the localization of the help pages

In reply to this post by sylvestre
2011/5/18 Sylvestre Ledru <[hidden email]>:

> Le samedi 14 mai 2011 à 23:13 +0200, Kenneth Nielsen a écrit :
>> Hey
>>
>> > Hi!
>> >
>> > It seems that it will be problematic to make the full translation during
>> > the next release cycle.
>> >
>> > On the other hand po2xml generates docbook only for 100% translated PO
>> > that is 100% XML-compliant.
>> > It is good for consistency, but Rosetta do not check the XML (it is well
>> > known that Ubuntu docs are broken for many languages because the
>> > translators ignore the guidelines or use the wrong guidelines). In KDE we
>> > use pology+Lokalize to debug POs with XML code, but it can be the overkill
>> > for Scilab.
>>
>> Alternatively you can also use gtxml (GetTextXML) from the pyg3t
>> package (of which I'm a co-author, this is really just shameless
>> adverticement ;)). gtxml is written by a fellow Danish translator Ask
>> H. Larsen. It is a commandline tool that generates helpful error
>> messages, so it can be used e.g. for checking the xml of all the
>> localised files and sending out a combined email with errors for all
>> of them.
> Do you have some documents and use cases of it ?

Unfortunately not. The only documentation at the time is what is in
the --help function:
Usage: gtxml [OPTION]... [FILE]...

Parse the contents of each po-FILE, writing warnings for entries suspected of
containing ill-formed xml.  A translated entry is considered ill-formed if its
msgid is well-formed xml while at least one of its msgstrs is not.  If no FILE
is given, or if FILE is -, read from stdin.

Options:
  --version      show program's version number and exit
  -h, --help     show this help message and exit
  -s, --summary  write only whether each FILE passes or fails, and the number
                 of valid and invalid strings for each file.
  -f, --fuzzy    print warnings for fuzzy entries aside from just translated
                 entries.

As for use cases I can provide you with some if you like, but as you
can see from the above it is pretty simple.

As I said earlier, pyg3t is still a relatively young project, so it
would be probably be ok to use for generating warnings that could be
sent to an email list, but I would probably be a little hesitant of
using it for build time checks at this point. At least you should
know, that if you do, you will be the first ones.

Regards Kenneth