[jitsi-dev] Slovenian missing from Pootle


#1

Hi,
please add Slovenian to Pootle so I can localize it.
Thanks, m.


#2

please add Slovenian to Pootle so I can localize it.

Done:
http://translate.jitsi.org/sl/sip_communicator/

Thanks in advance for your work!

Thanks, m.

Ingo


#3

Translating Jitsi is a tough job. A first translation takes a relatively big amount of time. As a result we have a number of started but incomplete translations that, when using them, actually mostly show up the English strings, but that still ship with the binaries.

This probably creates a false expectation and disappointment in a number of users.

I am wondering if we shouldn't instead only ship translations that can be considered complete. Let's say, any thing above 66%. This would take us to about 14 languages out of the 35 we currently have.

Thoughts?

Emil

P.S. We should probably also do a good clean up of the resource files and get rid of the tons of strings that aren't really used but that's a different topic.

···

On 14.07.13, 21:01, Ingo Bauersachs wrote:

please add Slovenian to Pootle so I can localize it.

Done:
http://translate.jitsi.org/sl/sip_communicator/

Thanks in advance for your work!

Thanks, m.

Ingo

_______________________________________________
dev mailing list
dev@jitsi.org
Unsubscribe instructions and other list options:
http://lists.jitsi.org/mailman/listinfo/dev

--
https://jitsi.org


#4

Translating Jitsi is a tough job. A first translation takes a relatively big amount of time. As a result we have a number of started but incomplete translations that, when using them, actually mostly show up the English strings, but that still ship with the binaries.

This probably creates a false expectation and disappointment in a number of users.

I am wondering if we shouldn't instead only ship translations that can be considered complete. Let's say, any thing above 66%. This would take us to about 14 languages out of the 35 we currently have.

Thoughts?

+1

Yana

···

On Jul 14, 2013, at 9:18 PM, Emil Ivov <emcho@jitsi.org> wrote:

Emil

P.S. We should probably also do a good clean up of the resource files and get rid of the tons of strings that aren't really used but that's a different topic.

On 14.07.13, 21:01, Ingo Bauersachs wrote:

please add Slovenian to Pootle so I can localize it.

Done:
http://translate.jitsi.org/sl/sip_communicator/

Thanks in advance for your work!

Thanks, m.

Ingo

_______________________________________________
dev mailing list
dev@jitsi.org
Unsubscribe instructions and other list options:
http://lists.jitsi.org/mailman/listinfo/dev

--
https://jitsi.org

_______________________________________________
dev mailing list
dev@jitsi.org
Unsubscribe instructions and other list options:
http://lists.jitsi.org/mailman/listinfo/dev


#5

Translating Jitsi is a tough job. A first translation takes a relatively
big amount of time. As a result we have a number of started but
incomplete translations that, when using them, actually mostly show up
the English strings, but that still ship with the binaries.

This probably creates a false expectation and disappointment in a number
of users.

Well, I always thought the same. But I opened the translations anyway in the
good hope that the requestors actually fill them in. This might be naïve.

I am wondering if we shouldn't instead only ship translations that can
be considered complete. Let's say, anything above 66%. This would take
us to about 14 languages out of the 35 we currently have.

I'm not sure if the best thing is to leave them out completely when they are
below 66% (or any number for the matter). But maybe we could show the
completeness just besides the language in the UI?

Thoughts?

Emil

P.S. We should probably also do a good clean up of the resource files
and get rid of the tons of strings that aren't really used but that's a
different topic.

Speaking of that: I recently did some research what to use to translate the
error messages of dnssecjava. As these are very simple strings I just used
ResourceBundles. But I came across a few interesting libraries that deal a
lot better with the peculiarities of "changing" strings in various
languages. E.g. "(He|she) edited N messages (one|M) second(s) ago" where He
or She would come from a gender enumeration and correctly deals with N and M
in languages with different plurals than the easy languages English, French
or German. As a goody, the strings are retrieved from a class that provides
methods for each key and takes the parameters as type-safe arguments.
To get rid of strings one could simply let any refactoring tool analyze
which method is not called anymore.

Thoughts?
(Yes, I understand that this would be a big thing.)


#6

I see two issues here. One is that some translations are incomplete, and there is no way to know in the binary if this is the case.

The other is that some translations are incomplete because they were not finished...

On the first, an additional line in the name (33% incomplete) does two things; warns the users, and advertises to bilingual people! On incomplete templates, an easy link to how to take over an abandoned template would also be helpful. Part of the problem is that people can not easily find out how to contribute. For example, where is the list of partial translations?

      Lee

···

On 07/14/2013 02:18 PM, Emil Ivov wrote:

Translating Jitsi is a tough job. A first translation takes a relatively
big amount of time. As a result we have a number of started but
incomplete translations that, when using them, actually mostly show up
the English strings, but that still ship with the binaries.

This probably creates a false expectation and disappointment in a number
of users.

I am wondering if we shouldn't instead only ship translations that can
be considered complete. Let's say, any thing above 66%. This would take
us to about 14 languages out of the 35 we currently have.


#7

Translating Jitsi is a tough job. A first translation takes a relatively
big amount of time. As a result we have a number of started but
incomplete translations that, when using them, actually mostly show up
the English strings, but that still ship with the binaries.

This probably creates a false expectation and disappointment in a number
of users.

Well, I always thought the same. But I opened the translations anyway in the
good hope that the requestors actually fill them in. This might be naïve.

Oh, please don't get me wrong. This was in no way criticism. If you or anyone else hadn't added those languages I would have done so myself. Anything else would be stifling contributions.

That's on Pootle though. Everything I am saying is about which of those we then ship in Jitsi.

I am wondering if we shouldn't instead only ship translations that can
be considered complete. Let's say, anything above 66%. This would take
us to about 14 languages out of the 35 we currently have.

I'm not sure if the best thing is to leave them out completely when they are
below 66% (or any number for the matter). But maybe we could show the
completeness just besides the language in the UI?

Or maybe we could do both. Impose a lower limit (e.g. translations with less than 33% are really pointless) and show those having less than 80% completion with a warning of some sort.

My suggestion about 66% was based on the fact that its something that we could do very easily either on the pootle server or in the installer build scripts. It will be as simple as running "wc -l" on the properties files.

Thoughts?

Emil

P.S. We should probably also do a good clean up of the resource files
and get rid of the tons of strings that aren't really used but that's a
different topic.

Speaking of that: I recently did some research what to use to translate the
error messages of dnssecjava. As these are very simple strings I just used
ResourceBundles. But I came across a few interesting libraries that deal a
lot better with the peculiarities of "changing" strings in various
languages. E.g. "(He|she) edited N messages (one|M) second(s) ago" where He
or She would come from a gender enumeration and correctly deals with N and M
in languages with different plurals than the easy languages English, French
or German. As a goody, the strings are retrieved from a class that provides
methods for each key and takes the parameters as type-safe arguments.
To get rid of strings one could simply let any refactoring tool analyze
which method is not called anymore.

Thoughts?

Sounds very interesting! I mostly like the fact that this makes use of strings more economical.

I wonder if it would do anything to help reduce the number in other ways. While not that easy, discovering entirely unused strings isn't that hard now and can be handled with a simple shell script. The hard part would be to detect strings that are used but never shown. Or strings that could be merged.

(Yes, I understand that this would be a big thing.)

Yeah :(. I don't see us (BJ) doing anything like this in the near future. We should still keep it in mind though.

Would it easily work with pootle by the way?

Emil

···

On 14.07.13, 21:29, Ingo Bauersachs wrote:

_______________________________________________
dev mailing list
dev@jitsi.org
Unsubscribe instructions and other list options:
http://lists.jitsi.org/mailman/listinfo/dev

--
https://jitsi.org


#8

I am wondering if we shouldn't instead only ship translations that can
be considered complete. Let's say, any thing above 66%. This would take
us to about 14 languages out of the 35 we currently have.

I see two issues here. One is that some translations are incomplete,
and there is no way to know in the binary if this is the case.

That would be a part of the build process.

The other is that some translations are incomplete because they were not
finished...

Well, there's not much difference between never finished and abandoned: both
simply lack strings and/or the correction of changes.

On the first, an additional line in the name (33% incomplete) does two
things; warns the users, and advertises to bilingual people! On
incomplete templates, an easy link to how to take over an abandoned
template would also be helpful. Part of the problem is that people can
not easily find out how to contribute. For example, where is the list
of partial translations?

Yes, a link to our Pootle (http://translate.jitsi.org where you can see
completion status, the uncompleted strings, etc.) in the configuration
dialog would probably encourage translators.

      Lee

Ingo


#9

I am wondering if we shouldn't instead only ship translations that can
be considered complete. Let's say, anything above 66%. This would take
us to about 14 languages out of the 35 we currently have.

I'm not sure if the best thing is to leave them out completely when
they are below 66% (or any number for the matter). But maybe we could
show the completeness just besides the language in the UI?

Or maybe we could do both. Impose a lower limit (e.g. translations with
less than 33% are really pointless) and show those having less than 80%
completion with a warning of some sort.

My suggestion about 66% was based on the fact that its something that we
could do very easily either on the pootle server or in the installer
build scripts. It will be as simple as running "wc -l" on the properties
files.

Sure, I don't care about the actual number.

Thoughts?

Emil

P.S. We should probably also do a good clean up of the resource files
and get rid of the tons of strings that aren't really used but that's a
different topic.

Speaking of that: I recently did some research what to use to translate
the error messages of dnssecjava. As these are very simple strings I
just used ResourceBundles. But I came across a few interesting
libraries that deal a lot better with the peculiarities of "changing"
strings in various languages. E.g. "(He|she) edited N messages (one|M)
second(s) ago" where He or She would come from a gender enumeration and
correctly deals with N and M in languages with different plurals than
the easy languages English, French or German. As a goody, the strings
are retrieved from a class that provides methods for each key and takes
the parameters as type-safe arguments. To get rid of strings one could
simply let any refactoring tool analyze which method is not called
anymore.

Thoughts?

Sounds very interesting! I mostly like the fact that this makes use of
strings more economical.

I wonder if it would do anything to help reduce the number in other
ways. While not that easy, discovering entirely unused strings isn't
that hard now and can be handled with a simple shell script. The hard
part would be to detect strings that are used but never shown. Or
strings that could be merged.

A first step would be to remove duplicates. E.g. the address book plugins
all duplicate the "Prefix" string. And if the shell script to remove
completely unused string is really so easy then please run it :slight_smile: I'm sure
there are about 20% of the string unused. But I'm also sure you don't get
them all because of the (you know my opinion) line wrapping at 80 chars.

(Yes, I understand that this would be a big thing.)

Yeah :(. I don't see us (BJ) doing anything like this in the near
future. We should still keep it in mind though.

You once talked about BJSoC. Or maybe it could be something as an internship
for an early-semester student of yours?

Would it easily work with pootle by the way?

Yes. The template format is a bit different though: {0:GENDER} instead of
just {0} or something the like. Depends on the library in the end.

Emil

Ingo


#10

So, looking at that list I saw some Spanish missing. I can help with that. I also saw two Portuguese... Is there some reason the missing parts can't use Brazil's version for now? With a little promotion, this is an easy problem to fix!

      Lee

···

On 07/15/2013 12:50 PM, Ingo Bauersachs wrote:

On the first, an additional line in the name (33% incomplete) does two
things; warns the users, and advertises to bilingual people! On
incomplete templates, an easy link to how to take over an abandoned
template would also be helpful. Part of the problem is that people can
not easily find out how to contribute. For example, where is the list
of partial translations?

Yes, a link to our Pootle (http://translate.jitsi.org where you can see
completion status, the uncompleted strings, etc.) in the configuration
dialog would probably encourage translators.


#11

So, looking at that list I saw some Spanish missing. I can help with
that.

Well, feel free to translate or correct what's missing or wrong.

I also saw two Portuguese... Is there some reason the missing
parts can't use Brazil's version for now? With a little promotion, this
is an easy problem to fix!

Be careful here: The Pootle server has two projects, Jitsi and Terminology.
The statistics on the homepage are the combination for both projects and is
irrelevant. The only one that matters is at [1] and there we don't even have
a Portuguese (Brazil) translation.

Now if you say that these stats shouldn't be shown, I absolutely agree with
you. However, I don't want to delete the Terminology project as it is used
for suggestions of words during translation, and hiding it is not possible
(or at least I haven't found an option during my long adventures in setting
this beast up).

      Lee

Ingo

[1] http://translate.jitsi.org/projects/sip_communicator/


#12

Looking closer, I am more confused... There are translations and suggestions. Which do I use? How do suggestions become translations? Is there a translation howto somewhere?

On that note, should we still be using the mailing lists as a table of contents? :wink:

      Lee

···

On 07/16/2013 01:17 AM, Ingo Bauersachs wrote:

So, looking at that list I saw some Spanish missing. I can help with
that.

Well, feel free to translate or correct what's missing or wrong.

I also saw two Portuguese... Is there some reason the missing
parts can't use Brazil's version for now? With a little promotion, this
is an easy problem to fix!

Be careful here: The Pootle server has two projects, Jitsi and Terminology.
The statistics on the homepage are the combination for both projects and is
irrelevant. The only one that matters is at [1] and there we don't even have
a Portuguese (Brazil) translation.


#13

Be careful here: The Pootle server has two projects, Jitsi and
Terminology. The statistics on the homepage are the combination for
both projects and is irrelevant. The only one that matters is at [1]
and there we don't even have a Portuguese (Brazil) translation.

Looking closer, I am more confused... There are translations and
suggestions. Which do I use?

To have Jitsi translated, you need to translate the project Jitsi (the
symbolic name is still sip_communicator).

How do suggestions become translations?

Never. Suggestions are shown while you're translating a project. While
editing a string, and if this string contains a word and that word is
translated in the terminology, a hint is shown that "button" in English is
"botón" in Spanish (if my dict is right).

At least that’s my understanding. Not sure if this is absolutely correct, I
never used suggestions for the German translation.

Is there a translation howto somewhere?

Specific for Jitsi, no. Pootle has its own documentation [1], section Online
translation editor [2] might be interesting for you.

On that note, should we still be using the mailing lists as a table of
contents? :wink:

I don't get that...

      Lee

Ingo


#14

A tongue in cheek comment about the requests to the list for instructions on pootle, or how to find a voip provider, or lists of know good providers, or how to provision...

We have a lot of documentation, and a LOT of very good resources, but they seem to be hard to find. :slight_smile: Do we have a web or documentation team? I would be very willing to join and help out there.

      Lee

···

On 07/16/2013 01:14 PM, Ingo Bauersachs wrote:

On that note, should we still be using the mailing lists as a table of
contents? :wink:

I don't get that...


#15

On that note, should we still be using the mailing lists as a table of
contents? :wink:

I don't get that...

A tongue in cheek comment about the requests to the list for
instructions on pootle, or how to find a voip provider, or lists of know
good providers, or how to provision...

Yannik wanted to collect a list of providers and contact them, I don't what
the status is on this?

Provisioning already has a relatively long manual, but it's probably well
hidden and could need an overhaul.

What I'd find important is restructuring the FAQ. E.g. a split-up in one for
beginners, advanced users and developers.

We have a lot of documentation, and a LOT of very good resources, but
they seem to be hard to find. :slight_smile: Do we have a web or documentation
team? I would be very willing to join and help out there.

Well, David Bolton did some great work on the user documentation page. Other
than that, there is no special team. Our website is based on PmWiki, so if
you have something in mind, you could just send it as a draft to the list
and we'll put it up.
After a couple of good articles, we can probably open an account for you to
edit it directly.

      Lee

Ingo

···

On 07/16/2013 01:14 PM, Ingo Bauersachs wrote:


#16

Hi guys,

I think it'd be useful to have a skeleton of desired topics for user docs.
Perhaps someone with some experience could draft out major topics (with a few minor topics in each) that should be covered and then it'd be easier to know what to draft up.
Also some guidelines on the format or preferred styles of submission, if there are any.
The video game framework Libgdx has something like this (link: https://code.google.com/p/libgdx/wiki/TableOfContents)

Just a thought.

Nick

···

From: ingo@jitsi.org
To: dev@jitsi.org
Date: Wed, 17 Jul 2013 18:53:12 +0200
Subject: Re: [jitsi-dev] List as a TOC ( Was: Filtering unfinished translations)

> On 07/16/2013 01:14 PM, Ingo Bauersachs wrote:
>>> On that note, should we still be using the mailing lists as a table of
>>> contents? :wink:
>>
>> I don't get that...
>
> A tongue in cheek comment about the requests to the list for
> instructions on pootle, or how to find a voip provider, or lists of know
> good providers, or how to provision...

Yannik wanted to collect a list of providers and contact them, I don't what
the status is on this?

Provisioning already has a relatively long manual, but it's probably well
hidden and could need an overhaul.

What I'd find important is restructuring the FAQ. E.g. a split-up in one for
beginners, advanced users and developers.

> We have a lot of documentation, and a LOT of very good resources, but
> they seem to be hard to find. :slight_smile: Do we have a web or documentation
> team? I would be very willing to join and help out there.

Well, David Bolton did some great work on the user documentation page. Other
than that, there is no special team. Our website is based on PmWiki, so if
you have something in mind, you could just send it as a draft to the list
and we'll put it up.
After a couple of good articles, we can probably open an account for you to
edit it directly.

> Lee

Ingo

_______________________________________________
dev mailing list
dev@jitsi.org
Unsubscribe instructions and other list options:
http://lists.jitsi.org/mailman/listinfo/dev


#17

I think it'd be useful to have a skeleton of desired topics for user
docs.

This is indeed a good idea!

Perhaps someone with some experience could draft out major topics
(with a few minor topics in each) that should be covered and then it'd
be easier to know what to draft up.

Anyone from the users to step out? :slight_smile:
It usually doesn't come out very good if developers write it (the best proof
is the current situation).

Also some guidelines on the format
or preferred styles of submission, if there are any.

I think David did a good job with the doc pages he created and we should
follow the style of these. There was a short discussion on the list to use
title casing [1], but that is pretty much the only rule we kind of defined.

As for submission, for the time being just send new pages to the list, and
requests for corrections e.g. in the form as David did before he had his own
account [2].

The video game
framework Libgdx has something like this (link:
https://code.google.com/p/libgdx/wiki/TableOfContents)

Just a thought.

Nick

Ingo

[1] http://lists.jitsi.org/pipermail/dev/2013-January/005653.html
[2] http://lists.jitsi.org/pipermail/dev/2013-January/000579.html


#18

Hi guys,

I think it'd be useful to have a skeleton of desired topics for user docs.
Perhaps someone with some experience could draft out major topics
(with a few minor topics in each) that should be covered and then it'd
be easier to know what to draft up.

Here's some missing pages that come to mind:

==Add a Network==
* XMPP Account (explain the term in ways that are directly relevant to
the user. Name examples of recommended XMPP services. Give instructions
for adding an XMPP account to Jitsi)
* SIP Account (should parallel info in XMPP article)

==Overview==
* Main Window (briefly describe each part of the main window, including
each area, button field and user information. It would probably make
sense to write this page last so that you can link to the detailed pages
for each topic)

==Contact Management==
* Add a Contact
* Merge Contacts
* Remove a Contact
* Create and Manage Groups

==Communicate==
* Text Chat (Including how to setup and verify encrypted chat)
* Group Chat
* Voice Call (Including how to verify encrypted chat)
* Conference Call
* Video Call (Including how to verify encrypted chat)
* Conference Video (possibly merge with Conference Call article above)
* Share Screen (including full screen, region, and remote control)
* Send a File
* Auto Answer and Call Forwarding

Eventually we'll need to document other things (such as the Options
window) but the above information would at least get new users started.

Here's a link to the current documentation:
https://jitsi.org/Documentation/UserDocumentation

Also some guidelines on the format or preferred styles of submission,
if there are any.
The video game framework Libgdx has something like this (link:
https://code.google.com/p/libgdx/wiki/TableOfContents)

In general strive for good documentation writing. To mention a few points:

* Use active voice
"Jitsi encrypts your communication..." rather than "Your communication
is encrypted..."

* Use 2nd person rather than 3rd person
"You can call..." rather than "One can call"
(People don't speak in the 3rd person except in academia. 2nd person is
usually easier to follow)

* Avoid blaming the user. In these situations you may need to use
passive voice.
"The letters and numbers do not match" rather than "You mistyped the
letters and numbers"

* Instructions should go from big to small
"From the main menu choose 'File > Add New Contact...'" rather than
"Choose 'Add New Contact...' from the 'File' menu"
(For the first sentence, the reader can follow the instructions step by
step. For the second sentence the reader has to mentally rearrange the
sentence in order to follow the instructions)

* When possible, explain relevant concepts clearly and succinctly
without resorting to jargon
"Facebook does not let you chat with people on other services such as
Google or Yahoo" rather than "Facebook Chat does not federate"
(For example it could take weeks or even months for a newcomer to
understanding the current FAQ without any background in XMPP, SIP,
specialized encryption terms, etc. However some jargon is unavoidable
due to the nature of the software (the words XMPP, SIP, and encryption
may be unavoidable due to the nature of the software).

···

On 7/17/2013 12:37 PM, Nick H wrote:


#19

Hi Ingo,

Good point I'll give a hack at it. I cooked up a little HTML for the smileys as I wanted to (haven't had a moment free to show it) but consider it something I'll be working on.

Thanks,
Nick

···

From: ingo@jitsi.org
To: dev@jitsi.org
Date: Wed, 17 Jul 2013 19:57:59 +0200
Subject: Re: [jitsi-dev] List as a TOC ( Was: Filtering unfinished translations)

> I think it'd be useful to have a skeleton of desired topics for user
> docs.

This is indeed a good idea!

> Perhaps someone with some experience could draft out major topics
> (with a few minor topics in each) that should be covered and then it'd
> be easier to know what to draft up.

Anyone from the users to step out? :slight_smile:
It usually doesn't come out very good if developers write it (the best proof
is the current situation).

> Also some guidelines on the format
> or preferred styles of submission, if there are any.

I think David did a good job with the doc pages he created and we should
follow the style of these. There was a short discussion on the list to use
title casing [1], but that is pretty much the only rule we kind of defined.

As for submission, for the time being just send new pages to the list, and
requests for corrections e.g. in the form as David did before he had his own
account [2].

> The video game
> framework Libgdx has something like this (link:
> https://code.google.com/p/libgdx/wiki/TableOfContents)
>
> Just a thought.
>
> Nick

Ingo

[1] http://lists.jitsi.org/pipermail/dev/2013-January/005653.html
[2] http://lists.jitsi.org/pipermail/dev/2013-January/000579.html

_______________________________________________
dev mailing list
dev@jitsi.org
Unsubscribe instructions and other list options:
http://lists.jitsi.org/mailman/listinfo/dev