Launchpad and Translation Process

[devnet]

Heya guys/gals!

I'd like to introduce you all to Og Maciel. He's a programmer (pet
projects Billreminder and PyLyglot), gnome developer, and a current
translator of XFCE, KDE, Openbox, and Gnome. He's also been in charge
of the entire Brazilian-Portuguese translation team for Ubuntu.
There isn't much that Og (pronounced Oggie) hasn't done when it comes
to translation.

I brought Og into our mailing list/group so that you all could ask him
questions and pick his brain so to speak. I know that he hosts and
has hosted his own translation projects for what he programs and I
also know that he's helped on large teams of people...I think his
experience is unique and could also help us out here.

So without further elaboration, please ask Og any questions you might
have....AND I'd like to ask that Og give us a couple of ideas on how
translation projects can organize and empower their members. Thanks
for helping Og! Please PCLOS i18n, ask him any questions you might
have!

devnet

[Ramboz Pierre-Henri]

So welcome aboard you both ...

2008/10/2 devnet <dev...@gmail.com>

[Terry McKenna]

Welcome Guys.

My first concern is that this project is not clearly ratified by Textar.

A project without the main stakeholder out of the picture questionable!

I'd like to see this project ratified by the Chief before going ahead. 

Terry

2008/10/3 Ramboz Pierre-Henri <did...@gmail.com>

[devnet]

On Thu, Oct 2, 2008 at 9:22 PM, Terry McKenna <mckenn...@gmail.com> wrote:
> Welcome Guys.
> My first concern is that this project is not clearly ratified by Textar.
> A project without the main stakeholder out of the picture questionable!
> I'd like to see this project ratified by the Chief before going ahead.
> Terry
>

This can be taken care of quite easily. I have a long working
relationship with Texstar and am confident that things can be more
than worked out. In fact, there have been many times in the past
where Tex has said he would like a multi-lingual DVD release. If we
can collaborate with Santa's Helper rpm rollers and this team, we
really could do something great after the next base release :)

[David Smid]

This is a copy of my reply from
http://www.mypclinuxos.com/forum/index.php?topic=2303.0
--------------------------------------------------------------------------------

Gettinther, I think you haven't read or didn't quite understand our discussion
at Google Groups. I must admit it's a little bit chaotic. So let me explain our
ideas and proposals here.

You are wondering why are not the translations sent upstream. Answer is easy:
there are close to zero translations that fit to upstream releases.

There are three groups of packages we maintain l10n for:

1. Packages that are not any specific to PCLOS and are not modified by any PCLOS
patches.
Such packages usually have good l10n (many of them come from Mandriva and
Mandriva is very i18n friendly). A agree with you, that we should not maintain
l10n for these packages and if, then only temporarily before our contribution
hits upstream. I think we do not have many translations of these packages in our
CVS/SVN. Example: kdebase, krusader, firefox

2. Packages that were born in PCLOS, made by PCLOS packagers and developers.
These should be definitely maintained by our i18n project. There are only few of
them but the count is growing.
Example: VIT, gtk-liveinstall

3. Packages that are taken from upstream but somehow modified to fit into PCLOS.
Such modifications usually also change the program messages and the l10n from
upstream is not sufficient for our needs. This is the majority of packages we
maintain l10n for. It's quite tricky - we must keep merging our translations
with upstream translations.
Example: drakxtools, synaptic, ...

What are our options to deliver our translations to users ?

1. Current situation: we can hope packagers will contact us and include our work
to their packages upon new release. This method is very unreliable.

2. After we somehow change the translations or after we add a new language, we
can contact packager and pursue him to add our modification and make a new
release. This is not the way to go: this project just doesn't have enough
manpower to negotiate with all the packagers all the time. Such administrative
would be very demanding. And what if someone refuses to include our translations
? Should we give up localizing that particular package ?

3. There could be a rule: no package will be put into testing if it doesn't
contain latest translations from i18n project. No exceptions. We could offer
packagers help with i18n, we could provide them with patterns, Makefile and
.spec snippets.

4. We could have (actually we already have) some translation release system to
easily:

* maintain translations together with information about package version
* merge our translations with upstream translations
* automate .desktop files translation
* build a (S)RPM with one command

This RPM would only contain translations and .desktop files - nothing else.
Never ever.
The RPM would be made in a such way that it will backup and replace any l10n
files already installed by the core RPM. We can use the "alternatives" facility
for that, together with triggers. I'm pretty sure it wouldn't break anything.
The main advantage would be that the packagers and developers wouldn't have to
care about l10n at all. Just internationalize your project - and we'll do the rest.

There is another advantage:
Minimizing network traffic. L10n packages would be very small and not all
packages would have its l10n counterpart (only packages of type 2 or 3). What if
the core package has 20MB and we've just added a Thai translation ? New release
would mean downloading 20MB for everyone, put it in extra l10n package and only
users interested in l10n would have to download few kilobytes on update.
OK, but what about repository metadata ? Again, only minority of packages would
have its l10n counterpart. L10n packages could have its own section that could
be disabled for English-speaking users by default.

5. The same option as point 4 but without official support from PCLOS devs. We
would have to maintain our own unofficial repository.

I prefer option 4 but would be happy with option 3 as well and would be very
unhappy with option 5.

I am not naive and I know that the decision is always yours and there's nothing
we can do about it. But I would be very glad to hear, that you are recognizing
us as the partners for discussion.

[devnet]

So no one has any questions for Og on this? No questions at all to
tap into his many years of experience on managing massive translation
projects or just to see what he thinks of our process here to give any
constructive criticism on how we can get more efficient?

[Ramboz Pierre-Henri]

hi

I think I will speak for everyone in the following message.

Og is registered to this list so he may introduce his point of view like you do devnet. Since Og has much more experience than us all on l10n we are looking forward to read from him... he's the new one there, we never read a line from him... is he really interested in collaborating on this project or has he been committed to this list to be our all mighty mentor ?

There is no hanger but yet sarcasm in this reply but i think it depicts quite properly what is the feeling instilled from all this week new event.

1 : our work has been buried with no warnings
2 : our work has been excavated with a new toolset but no team
3 : team members have been informed by forwarded mails about the two previous fact on the mailing list
4 : we all discovered the "future" of our project was discussed by a few on a 3rd party site, independent (yet friend and partner) of our regular means of communication
5 : we've been granted of a new consultant coming with "many years of experience on managing massive translation
projects".
6 : and we are expected to ask questions when our "situation" isn't even commented !

Please Og ... if you want to comment something, please do with David message which states quite properly what is our situation and why the project is stalled but not dead, yet today two tools and their translation got updated by gianvaca ! Tell us your opinion !

Devnet, please, if you want to help, do so...
But at the very moment, i think you're the one who is seeking answers so ask us questions... i think you got many answers yet ... if you think we should use launchpad, rethink the amount of "administration it require for us to deal with and the amount of people we are ... put both in balance ... and consider that something discussed a year ago, with a diminution of 50% inb team members can be discussed again but won't reduce the power wasted in administration and not put in translation.

Now i think you understood what is our problem... we just need a mean of "distribution" ... we are able to do solution 5 but is it a good method ? The commitment of texstar on i18n package has always been verbal and is not (as for now) a feature that the development process of pclinuxos is endeavouring. So despite the allegation of i18n content in 2008... i must admit I'm sceptical on this very achievement and I would like to know who is the french translator of the new PCC?

Now, speaking for me !

I stopped my translations for some very simple reason (among others) Since our translation have only seldom been put in packages updates, Texstar and the gang are aware of our work, no one (in the gang) notifies us of update on i18n content or request patches and since most i18n patches have been applied by us in the current repo (or Santa team). Since only few of us are able to deal with complex srpm and handle multiple sources and since we don't have all access to the test repo. Since we have no feedback or comment on our work from the gang. Since we are frustrated by all that. WHY WOULD I WASTE MY TIME ?

So my last point will be to mention that I'm all for proposal 5 in David's plan. We are yet ready, able and willing to use this method to "hack" the current distro. . Our concern is not pclinuxos and the gang, our concern is the "average Joe" the "basic" user. Such a project will definitely end up in a remaster featuring one or more non English language... Will it be a fork? If we need support, feedback, "constructive criticism on how we can get more efficient" ... its on those points and if you feel like there is something to be said, please do so. but don't ask us to ask questions when we have wasted 1.5 year doing so without a single reply! (not from you both of course)

Devnet, Og, its up to you?
We do have a diplomatic matter here and we all know that some of you have experience in the domain and some other have the ability to open doors.

Once again, there is no hate or hanger in my mouth, just a lot of very "destructive" statement have been made on that project without taking the time to consider it as it is and not as it is not on mypclinuxos forums !

Thanks for your attention.

DidouPh aka Pierre-Henri RAMBOZ
ex pclos-i18n team leader

2008/10/8 devnet <dev...@gmail.com>

[Zoltan Hoppar]

Hi,

Well, I (we the HU team) think Didouph has right. We have been worked, did localized remasters, because of no one of the dev team gave a little attention - to "okay guys, lets put your work to mainstream". Yeah, 1.5 year has went away, without a promise, that we in i18n teams work are important part of the next release. Or somekind so featured and mentioned "internationalized DVD"... So, it was a wasted time or not?

Zoltan

2008/10/8 Ramboz Pierre-Henri <did...@gmail.com>

[Alessio Adamo]

I also understand the frustration of DidouPh and I partially share it. But I'm still optimistic, maybe because I joined the team only half a year ago. I've always been quite an active member, though I had often this feeling to work hard for no reason. I started without knowing anything about gettext, intltool, bash scripts and rpms but with some good will and help and patience from these guys (especially David, whom I consider my mentor) and the Santa's helpers, I became able to implement localization into bash scripts and into rpms by mastering gettext and intltool.
So far, my efforts were especially driven by mere thirst of knowledge, but now I'd like they can have some good use. Until we don't get a clear word from the developers, every effort in making this distro international, must be considered simply as an exercise. In these conditions, gathering new translation forces would be counter-productive, because uncertainty can quench and estinguish even the strongest enthusiasm.
I'm optimistic because Gettinther showed an interest into our project, and promised to discuss with us its operational organization once more urgent aspects of the 2008 distro are sorted but it's certainly useless to organize a plan of attack before this day.

PS. I still consider DidouPh and David as my project leaders.

2008/10/9 Zoltan Hoppar <hop...@gmail.com>

[Terry McKenna]

Show me post here by Textar with his vision spelled out, then I'll give you my thoughts/questions.

Terry

2008/10/8 devnet <dev...@gmail.com>

[Ramboz Pierre-Henri]

@ Gianvaca :

Both gettinther and devnet are strong support for us ... the first made the translation of synaptic possible and also the one who FORCED some of our translation into main stream. The second was our first host and the very first one to rise our voice to the ripers ears!

@ terry

http://www.linux.com/articles/62199
http://pclosmag.com/html/global.html
http://www.pclinuxos.com/index.php?option=com_smf&Itemid=26&topic=43589.msg334586#msg334586 (n°16)

and google you'll find more

also browse pclinuxos.com forums !

Abut his position on our work, its pretty hard to say since we didn't get REAL feedback from him... i could post pms about his l10n DVD but that would not be valid.

To ALL :

1 : Tom Kelly (previous LiveCD project manager) proposed to interact between the Gang and i on the I10n project issues and livecd but i don't think its an option.

As you know i also lead the LiveCD project wich has been definately "absorbed" by the gang. They just informed me that they will "no longuer synch with the original repo" so at the very moment... LiveCD is no longuer a multiplatform project but a PCLinuxOS centric tool !

I such action and derive are dangerous and explain a lot of my scepticism.

2 : here is an extract of a chat devnet and i had yesterday about my previous message :

devnet: ya, I kinda get where you're coming from

22:20 I thought people might have questions about launchpad though

 moi: launchpad is the least of our concern

  honest

  if it ain't broken ... don't fix it !

 devnet: so what's the real problem?

22:21 moi: silence

 devnet: what I've heard is that no one has status

  we don't know what status of various translation is

 moi: status of all translation is "freaking ignored"

  by the gang

  all our work is sent

 devnet: ok, so this is an upstream problem

 moi: little is included

  exactly

22:22 devnet: the only way something will be done is if I force my way onto the dev team again

 moi: so all team go on a local remaster wich is trashed by every single repo update

 devnet: to be avoice for the community again

 moi: no

  the best way is to play the same game and open a 3rd party repo

  just ignore

 devnet: I don't think that would be good

 moi: why ?

22:23 devnet: I think that upstream cannot give feedback because they don't track anything

  I think they need systems in place to do that

 moi: honest

 devnet: and they don't have them

 moi: i don't think they simply CARE

22:24 devnet: I don't believe that

  because they're still working on it

 moi: when i read announces about the l10n DVD

  it seriously makes me laugh

 devnet: you have a link?

 moi: last anounce was on the interview @ distrowatch

22:25 and some on the forum

  and also a little hello by tex in 2007 to tell us @ mypclinuxos that he will come and give us hints on what was needed by the team for i18n content

  all translators left because of misstrust in the gang

22:26 devnet: ok...so here's what should happen

  the 3rd party repo should happen if my attempt at diplomacy and project management for the dev team fails (which is a large possibility)

  a 3rd party repo should happen if they fail to meet their obligation

  so we should develop side by side a third party repo

  but not announce it

22:27 moi: they yet failed

 devnet: of course

  but they can make accusations at this point in time

  we make sure we document our attempts

 moi: we don't ask for our name to be writen in fire on top of every single cd, we just want to be informed of update and being given the minimum to provide data

 devnet: we cover our asses

 moi: i yet have my mail database :p

22:28 devnet: let's make sure everyone sees us making these requests...publicly

so should we move forward on it and just let them come to us to use it?

 moi: its not a fork but a severe rise from the non anglo-saxo community

22:30 i think that the option you present (diplomacy) must be held again but with no excessive entousiasm

  for i do believe it will end up in an horrible failure

 devnet: Ok

 moi: but

22:31 i think its the good option

  and david convinced me yet that this is the thing to do

  but

  1 : i don't think that this should be discussed here but on the list

22:32 devnet: the list is public right?

 moi: 2 : that if we call to the dev its nothing to be done trough your voice but the team's one

  the list is public to read

 devnet: ok, let me push hard on them...it will fail

  we'll document it publicly

 moi: not you

  but we

22:33 devnet: ok, I was going to phone Tex

  :)

 moi: You , I or Og should not be taken as personification of anthing and neither should tex be

  lol

22:34 hehe

We all know that the only mistake with this project was to believe it could be indépendent and unofficial...

@ Texstar

I just don't want to go back to my 2006 winter and the release of a remastered pclinuxos cd every 15 days.

Please give me one good reason. Just one not to spend all my time and mone in releasing 3rd party DVD and RPM !
If the gang has no time or interrest, we are 20+ persons willing to do the job and ready to complie with the rest of the team rule... we don't ask for inclusion, we asc for information... just tell us what you're up to and we will translate it.

DidouPh

[Alessio Adamo]

No disrespect for devnet and Gettinther was meant of course, I just don't want you two, the people I mainly dealt with for this project, to throw in the sponge yet.
Let's wait until the beta is out to see whether someone will pat on our shoulder and ask for our skills.

2008/10/9 Ramboz Pierre-Henri <did...@gmail.com>