Re: Arabeyes Translation Teams

[Nadim Shaikli <shaikli at yahoo dot com>]

Before I start, let me note a couple of things ;-)

 1. What is noted is my personal opinion and does NOT constitute
    Arabeyes' official stance or direction or decision or ...
 2. I am thrilled to see lots of people chiming-in with thoughts
    and opinions and most importantly I am genuinely ecstatic to
    see that people continue to care - keep at it...
 3. Do please cite/include only relevant sections in your email
    replies instead of simply adding on to the entire previous
    email (keep in mind that our mail archiver threads - so no
    worries there).

--- Djihed Afifi <djihedlists at googlemail dot com> wrote:
> The more I look into the men count that we have for translation, the
> more I realise we are in a sore need for more translators. For

This is not simply a problem in translation - this is a problem with
almost everything we touch.  We have plenty of ideas yet time is
critical for most of us and there simply aren't enough helping or
interested hands.  With that said, I'm not citing an excuse, simply
pointing that this problem exists across all disciplines even beyond
open source efforts.  The reason I'm opting to stop and point out the
obvious will become apparent in a bit but in short the solution is
not entirely specific to the problem.

> example, for gnome 2.16, which is expected to be completed somewhere
> at the start of September, I'm seeing only three dedicated
> translators, the maintainer who is too stretched out and still doing a
> heck of a job (thank you Youssef), Raffah and myself. Let me put it
> simply to you: Arabic Gnome 2.16 won't ship with only us, I've thought
> about this realistically for a long time. Translation projects are
> slipping from under our attention quickly. Thankfully, we just barely
> have enough time to do something about it.

Here is a old idea that is worth repeating.  Why not have everyone that
is interested in translation (that is willing to work on something other
than their main project) work on ONE project until it is complete ?  In
other words, inform everyone that for the month of August ALL are kindly
asked to only contribute and translate the Gnome project, September
will get assigned to KDE, etc, etc.  There seems to be merit in such an
idea yet I'll leave it to those that do the actual work to decide.  With
the small'ish number of people that are doing the actual work this might
actually get a number of things accomplished sooner rather than later and
it will ultimately be more fun as there is more interaction and like-thought.

> Please take this seriously, we took it upon ourselves to be The
> Ones(tm) when it came to Arabic translation, so we should live up to
> the responsibility, Arabeyes is always here, and will never go under.
> The actual task: the translation should be the end goal, not the
> popularity of a website or the dedication of a handful of team
> members. The current system is too high for many potential
> translators, we're not gaining anything from it, we're actually
> locking people out of the translation effort in the name of
> "dedication" and "commitment", a goal Arabeyes has not even achieved
> beyond a couple dozen or so part-time translators. We need to make
> things *work*, not to over analyse people's behaviour philosophically.
> May be I'm too much of a pragmatic person, I believe pragmatism is
> what we need ATM and ASAP.

Yes and no.  What you see in terms of "complexity" or high barrier
for entry was the result of about 3-4 years of people like you doing
this work and trying to document it as best as they can.  In other words,
people wanted to streamline the process so that they don't continue to
get bombarded with questions, howtos, what-to-do's, etc.  Initially the
idea was for ONE project coordinator to be responsible for a project.
So say person-X comes around and he's interested in KDE.  He's really
really really interested and so he actually spends the time needed to
learn about CVS, etc and applies to be the project's maintainer.  That
person-X would then apply to be KDE's maintainer/coordinator and if
all pans out, he gets the assignment.  With that the role of 'Arabeyes'
proper ends in that person-X is now fully responsible for KDE (that
would include the files, the news, the sub-project page, blurbs, informing
people of tasks and schedules, keeping people motivated, recruiting, etc).
If this person-X after 2 months comes back and says "hey person-Z is
really good and he's helping, please give him an account" the 'core'
team would gladly do that.  There is _nothing_ stopping this person
from getting all the files translated on Neptune by martians which
email him the files back (without them knowing anything about Arabeyes
or CVS or ...) and he simply commits.  If person-X wanted to lock 
him/her-self in a dungeon for a month and do the entire work (post
telling the list that he'll vanish for that period of time) and then
simply surface with all the files completed so be it.

So why am I saying all of this ?  Well, that is because that is what
we started with.  Arabeyes' initial idea was just that (with relation
to translation).  As the projects grew and as people came-and-went
we started adapting to what is easier to maintain and how best to make
this more of a "process".  So it ultimately became easier to simply
give out CVS accounts to those that knew (or were OK'ed by someone
trusted to know) what they were doing.  Does that mean that they
can't contact our friends in Neptune for help ?  No it does not, that
option was and will continue to be available.  Meaning (and I'm jumping
the gun abit in talking about pootle, etc) people are free to use
and try any means they see fit to get the job done.  Arabeyes has
tried to be very accommodating yet with the unfortunate lull (and lack
of interest) on the 'core' side of things being accommodating now is
not something the handicapped 'core' members can realistically do.

Project maintainers/coordinators really need to step-forward and take-on
more of the burden to really __lead__ the project they are responsible
for since the more I think about our current state the more I'm inclined
to suggest the initial approach.

> This is also necessary to *contain* all Arabic translation efforts.
> Instead of being selective by raising the bar high for entrants,
> Arabeyes should embrace every possible translator. I wonder why we
> complain about not too many active members when we're being elitist.
> Again, the goal is not to form a group of elitists, let's drop the
> arrogance attitude a bit towards new comers, and move the Arabic
> translation train into its rails again. We're the only serious Arabic
> translators in the scene, it's a great *responsibility*, even
> religiously, if religion holds any place in you.

I don't think there is _any_ 'arrogance', I really don't.  But I can
guarantee you no matter what you'll do you won't get the level of
participation you'll expect.  The reason for this is simple (and I
don't mean anything negative by this but it is what veteran translators
have noted themselves) the job is boring and is never-ending.  Simply
put, no matter what you'll do you won't be able to retain a motivated
results-oriented team unless other carrots were introduced (money,
pride in ethnicity, competition versus something/someone else, etc)
and even that is not entirely sustainable.  As noted, this might seem
a bit strange for me to say (and I'd be the first one to note that)
but do ping the likes of Arafat, Abdulaziz, M.Elzubeir, A.AlRasheedan,
etc, etc.

So should we simply fold and give-up ?  Absolutely not.  I personally
think we are working on the wrong things when it comes to translation
in that we are attacking things with shear muscle and not brains.
If we have a complete word-list, a complete dictionary, a complete
thesaurus we could spend some time on an automatic translator - granted
it will suck in the beginning but that in my humble opinion is a
worth-while effort since the lists noted are bounded (once they are
complete, they are _done_ -- no need to repeat the entire process again
which needs to happen with every release of KDE, Gnome, OOo, etc).
Automatic translation also works well in say 70-80% in the strings we
have (Kbable and it's fuzzy translations are proof of that) which will
leave us say with 30% of what we currently hold to be revisited by-hand.
In short, our time collectively should be spent on laying the ground-work
on getting a system in place that will translate for us instead of us
being overwhelmed with the thousands of strings awaiting our attention.
Let's rethink what needs to happen not in terms of semantics and
mechanisms/mannerisms but in how can we revolutionize things...

> Please understand that 99.99% of Arabic computer users out there are
> not as tech savvy as you are (actually, that's the reason only you
> survived the obstacles..), the language(English) and the terminology
> (IT) form a learning curve that is simply too steep for them.

Linux (and open source) in western markets even is still a niche, so
I don't expect too much from the Arab computer user.  This is not a
question of savviness, but a question of interest, time and curiosity.
Most of those reading this (savvy folk :-) were simply curious about
linux and open source and started reading, using, trying and/or simply
asking about all of this.  In the process some who were inclined to
spent hobby (ie. lots) hours have been trying to improve on what is
out there when it comes to Arabic support - and along we all met :-)

Keep in mind that Arabeyes is not simply a translation effort but is a
development project in that we wanted to fix/patch that which we saw
broken and augment that which we saw missing (we tried to make the
procedure for both translation and development tracks similar to
simply things).  Development, unfortunately (if you will), will require
certain skill to accomplish and that's why we see a bigger hit on the
'developer' list, for instance, when it comes to addressing some of
the needs there.  In short, to develop you'll need to have this
'savvy' while in translation the project coordinator/maintainer could
easily shield people from having it given people know how to email
and abide by a minimal set of rules.

So to overly simply the entire process, a newbie sends an email
(in private if you will) to the project's maintainer to say "I want
to help" and the project maintainer sends him/her files to translate
while tracking progress and ownership.

> May be you're wondering what I'm expecting from this email. I'm
> expecting a revision of the whole translation process. Currently, new
> translators have to register to the website, read 99 manuals,
> subscribe to the mailing lists, get a CVS account (which means they
> have to be CLI masters). Remember, it's like a filter, every step in
> the way potentially discourages many people away. You require that
> level of dedication from may be core developers, or at most all
> developers, but not of every translator whose only interest is
> translating strings. After all, getting a CVS account and subs'ing to
> a mailing list is no guarantee for dedication, how many people have
> CVS accounts but are not or were never dedicated?

Agreed yet it was never meant to be a 'filter'.  Will it be difficult
for some ? Sure...  But again there is _nothing_ precluding people
from doing it their(tm) way.  Ask your cousins, the people next door,
your school principle, etc to all translate (using CP-1256 even :-)
and then have them all email you back the files (while you note to
those looking to help that the files are being worked on) and then
you can simply 'iconv' back to utf-8 and commit the files.  Feel free
to be as innovative as you like - if a method/system works and its
proven to really really really help then by all means tell us and
we'll try to get others to rally around the idea, but testing every
method out there is a bit cumbersome to say the least for the limited
people that are currently active.  If a tool becomes available that
everyone would like to seriously use (not just test and play around
with to see how we can modify it to suit our needs) then by all means
we'll try our best to get it up-n-running.

> Please understand that this is not a rant, I'll put my hand first on
> the table to get this moving. We need to *work*, to get things
> *working* and to show that we actually *mean business*.
> 
> Finally, after we fix our wrecked (in my opinion) translation and
> recruitment process, we should then think of real concrete steps to
> recruit more translators, I'll leave that after this discussion.
> 
> Please voice your opinion even if you'll only write a couple of words.

I too, like the others, agree but I think a different tack needs to be
taken.  My remark regarding "solution is not entirely specific to the
problem" steams from the fact I see this problem very related to what
is happening on alots of fronts beyond arabeyes - the mindset simply
needs to start changing and its not a question of mechanisms or tools
(I think its inherit within people - but I might be getting philosophical).
In other words, I don't think you'll get much accomplished/improved if
you were to remove all these "filters" (though as noted they potentially
aren't there given the coordinator shields them away); so prove me wrong :-)

In short, what you are asking for is semi-there.  You want a simplified
means to have newbies contribute and my response to that is "it is
already there" - you are welcome to do or try what you think will work.
All that is asked from you (given you a project maintainer/coordinator)
is that you keep CVS updated (sync'ed and committed with newly translated
files).  If you are looking to get pootle installed, I read a couple of
offers from people to do it for you to try.  My view on pootle (and ilk)
is that the idea is good but it is not fully cooked yet.  QaMoose, if you
really look into what it does, was/is a precursor to what pootle is
supposed to do - we stopped using it the way it was supposed to be used
due to lack of interest from supervisors who were supposed to inspect and
decide on (approve/reject) terms (Ossama.K care to comment on your
experience there ?).

So where does that leave us ?  Well, after such a long email I don't
really know - I forgot :-)  Seriously though, I can suggest a couple
of things (as noted above),

 1. All translators (as agreed among themselves) work on ONE project
    at a time.
 2. All translators work on the [technical] dictionary, wordlist, etc
    while rallying support and developers to get an automated system
    (auto-translator) in-place ASAP.

In all of the above I don't mean to shoot down any ideas or sway people
from doing one thing or another.  I'm trying to be as realistic as possible
while taking into account what I've seen in the past with relation to
various translation efforts as well as what is happening in the present
in hopes of providing some avenue for forward movement.

Just a thought (or two :-)

Salam.

 - Nadim



__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com