Seeking advice on team issues

[Roozbeh Pournader <roozbeh farsiweb info>]

I am writing this email because of my frustration with a situation in
the Persian GNOME team. I am seeking the advice of other people in the
GNOME project.

The situation is getting on my mind, specially because a dear friend and
colleague, Behdad Esfahbod, is on the other side of matter. It recently
got much worth because he checked in a translation by himself to the
GNOME CVS, without confirmation from me or other reviewers of the
Persian team (that at the same time removed a copyright line, which made
me more frustrated).

I can't speak for Behdad, but I can explain some of the processes that
has led to his objections, according to my interpretation of things.

To understand these, you should know that I became the Persian
translation coordinator only because I was the first translator (and the
only translator for a while), not because I have been elected or
anything. But nobody has challenged my role either, as far as I can
tell.

You should also know that Iran is not a signatory of any of the
international copyright conventions, but has an national copyright law
only protecting works first created in Iran.

1) Originally in the project, because of my worries over the copyright
status of things (which was mostly based on my sad experience with the
Persian KDE project) and my incomplete (but improving) understanding of
the copyright law, we had random policies imposed by me on the project. 

I used FSF as the copyright holder. This is the first Persian
translation ever checked in to the GNOME CVS:

http://cvs.gnome.org/viewcvs/gtk%2B/po/fa.po?rev=1.1

Later, I decided to ask (require?) every volunteer translator to sign
the famous FSF copyright disclaimer from the Translation Project and
send it to the FSF offices in paper before I check in any translation
from her or him to the GNOME CVS.

We used FSF copyrights for a while, but after a while found that
according to Iranian law, FSF will most probably be unable to use the
protection of the Iranian copyright law, which would make the work
practically public domain. I then decided to switch the copyright to the
FarsiWeb project, which was then a non-profit project by Sharif
University of Technology and Science and Arts Foundation. This was not a
problem since again almost all contributors were members of the FarsiWeb
project. We kept pointing project outsiders to the FSF disclaimer, but
changed the copyright of the existing files to the FarsiWeb project in
new CVS check-ins:

http://cvs.gnome.org/viewcvs/gtk%2B/po/fa.po?rev=1.26

Later, we found that since that project is no legal entity and cannot be
well-defined, it is not entitled to hold copyright over anything. At the
same time, the university and the foundation had pushed the FarsiWeb
project to become a spin-off company, which we founded in September 2003
(Behdad is a co-founder). We then transferred of rights to work we had
created to this company, and we started to use the company copyright. We
assumed that we can keep the work copyrighted in both Iran and outside
Iran by using a double copyright line, one for the company and one for
FSF:

http://cvs.gnome.org/viewcvs/gtk%2B/po/fa.po?rev=1.39

After a while, I found ways to recruit people to do GNOME translation,
including hiring a professional translator, Meelad Zakaria, to work on
GNOME translation. After a period of his working on GNOME, he was
becoming frustrated so we started to hold by-invitation translation
parties in Sharif University campus. It was a great success and has led
to activation of possible contributors (or them understanding that
Persian l10n is not really that easy). During these parties, several
friends and almost all company-employees contributed translations, which
were later reviewed and checked in to the GNOME CVS. It was with these
parties that we finally passed the first 50% frontier. Every person
participating in these translation parties that was not a company
employee has signed a copyright disclaimer in Persian, with a similar
text to the FSF translation disclaimer, but additionally explicitly
letting us transfer the rights to the FSF or any other organization we
feel appropriate. Like the FSF disclaimer, we mentioned that this
applies to translations of "freely redistributable software programs".

In my reading of the FSF disclaimer, which I have personally signed, it
applies only to translations that we provide to the FSF itself (i.e.
putting an FSF copyright notice on its header), but Behdad has claimed
that we the people who have signed that disclaimer "CANNOT claim
copyright" on any of our translations anymore. I am assuming that he
claims every translation of a free software program I will do in my life
after signing that disclaimer will be copyrighted by the FSF. But I
guess I may not speak for him. An email of his about the matter is here:

http://lists.sharif.edu/pipermail/farsi/2005-April/000328.html

But all this has an allegedly dark side. People who contacted us over
the Internet or did not reside in Iran had a harder time contributing to
the Persian translation effort than those working in the same room with
me or those participating in our translation parties. We somehow needed
to establish trust over email, which was not that easy.

Some people have wanted to impose their own Persian idiolect in their
translations and were offended when their translation were reviewed,
some of them considered that an attack against their personality. Others
left the project during times of inactivity, when some of their emails
were left unanswered or when I was foolishly waiting for a glossary from
the Persian Academy. Even Behdad has not yet passed the process of the
review of his work on gucharmap  (more about this later), which was done
at a time that he had moved outside Iran.

In short, only one contributor outside of FarsiWeb and translation
parties has had his work accepted for the GNOME Persian translation,
Hedayat Vatankhah:

http://cvs.gnome.org/viewcvs/gconf-editor/po/fa.po?rev=HEAD

That is in comparison to about twenty contributors whose work has been
checked in to the GNOME CVS, reviews of which was mostly done through
communications by email or in person or in translation parties.

Behdad objects to the current system, and claims that I have discouraged
some very good potential contributors by acting in certain ways. He has
also suggested that I am trying to create a copyright monopoly for the
company using these explanations as an excuse, by forcing contributors
to disclaim their rights in Iran or assigning them to the company.

He has also mentioned that me and the two other most active contributors
to the Persian translation effort have formed a conspiracy. I assume
that he believes it is a conspiracy that is helping me to do these wrong
doings. These two active contributors are Elanz Sarbar (my wife) and
Meelad Zakaria (a friend). Both are colleagues and FarsiWeb employees.

That was one story line. I appreciate any kind of advice about that.

2) The other is about a mailing list. We have not had an official
mailing list for the GNOME Persian localization project. There has been
a *closed* mailing list with open archives that was historically used
for communication between some selected people interested in Persian
localization, mostly concentrating on KDE at its early days. Since the
list resided on my server, and I did not see a need for a separate GNOME
Persian list, it has also been used for some of the communications
related to the GNOME translation into Persian. It has also been used as
an address for the Persian team in the Language-Team headers of the PO
files, but not the GTP Teams page or the page linked from it. The list
currently contains about one-third of the GNOME Persian contributors,
and even includes people who have (to my knowledge) never contributed
Persian translations to free software.

A while ago, for some reason, I removed an email from Behdad from the
archives of that mailing list, which resided on a machine for which I am
responsible to Sharif University for. While I don't consider that list a
part of the GNOME project, Behdad has mentioned that he supposes that it
is implied from a Language-Team header in GNOME Persian translations
that this is a mailing list of the GNOME Project, and that "The act of
removing messages from the archive for personal reasons is not tolerated
in the GNOME project."

I am seeking your advice again, specially if there is such a policy
about not removing emails from archives in GNOME, that a translation
team must have a mailing list, or that an email address mentioned at the
Language-Team header would count as a part of the GNOME project.

3) The third matter is about internal procedures and policies. In short,
the team's current practice has been to translate verbs that appear on
buttons not to their imperative forms in Persian, but to either
infinitive or passive forms. In my opinion, that is the current practice
in Persian software, and had been devised mostly because the users in
the Iranian culture would assume commands printed in buttons would be
towards them, not by them to the computer. In my opinion, that is also
because Persian is a more "polite" language, more like German than
English, which means that a singular form may look very informal but a
plural form may be confusing sometimes.

So it has been our existing policy to do that this way, and all the
checked in translations have either used the infinitive form or the
passive form, based on the context.

Behdad, in his first translation which he wanted to contribute to the
Persian GNOME, tried to change that practice. This was after he thought
he may check in his updates to the Persian translations without any kind
of review to the GNOME CVS, which I then explained to him and reverted.

The discussion about his suggested change was fruitful to some degree
and was joined by others, but we could not reach any kind of agreement.
I failed to agree with his points, and he insisted on his opinion. The
discussion was discontinued without reaching any kind of conclusion, and
the gucharmap translation which he has provided was not checked in.

This was before the translation parties and the new much more active
shape of things. During the parties, we did not assign gucharmap to
anybody, and Meelad emailed Behdad a few times with a request to submit
his latest work, so that it could be more informally reviewed so we can
get his work into GNOME.

During a time that we had Internet connection problems on 2005-11-05, he
emailed me and Meelad personally and said:

"I finished the gucharmap translation based on all the responses
on the list in early 2005.  I believe it should be committed as
is.  Any further discussion should go to the list.  Since
tarballs are due in a few hours, I will go on and commit before
making a tarball if it's not already committed."

That did not include the change we had requested for the verb on the
"Find" button, was removing a copyright notice, and may also have other
problems (but I have not checked it in detail yet). But before we had a
chance to review his work (and before we were even able to read his
email), he went and checked that in to the GNOME CVS about 33 hours
later, without my confirmation or any other active reviewer's.

I don't know if there is a GTP policy against this, but the current
policy of the Persian team (originally imposed by myself), does not
allow such acts.

Later, when Meelad reviewed Behdad's work, I went and committed that to
the CVS immediately (although it could have benefited from a second
review by me, I did not consider that necessary), and since I thought
that that translation which removed a copyright line was included in the
GNOME 2.12 release (which could have been the only reason Behdad had
committed the file without any review), went and emailed Noah and the
release team to re-release gucharmap with the new translation (that
included the copyright line and a fix to the disputed message, among
other things). But Behdad reverted my commit (which contained several
improvements to his translations), only restoring the removed copyright
line:

http://cvs.gnome.org/viewcvs/gucharmap/po/fa.po

I did not touch that again, in fear of getting this into a 

He has mentioned that "I have been patient enough to
read and follow tens of email communication, all for contributing
translations to gucharmap.  In return, I expect that any change
to the translations of that module pass through the same process."

I have two problems with that: first of all, he has only been patient,
but not accepting some of the comments, and second, I don't assume that
a process that has proved to be frustrating should be repeated for
updating his translation.

While I understand that he is now a co-maintainer of gucharmap, I don't
consider him having the final say about Persian translation matters in
that package, and I don't consider he has a right to revert
contributions or impose policies.

That was the third story. And the most important point for which I wish
to ask your advice. What would you recommend? Do you think somebody with
commit rights to the GNOME CVS (or a package co-maintainer for the
matter) has the right to commit his translations to GNOME without (and
against) the approval of the coordinator or the translation team?

The final notes:

Behdad had warned me in a personal communication (which I am reproducing
without his explicit permission) that:

"This is my last warning to you, the very next time I see you
making rules of yourself or anything when it comes to Persian
translation of GNOME, I don't wait a minute.  You know what I
mean."

In our previous private communications, he has mentioned his frustration
over the issues I mentioned above, and I had promised to try to solve
them, and has provided a deadline for me to try to solve them by the
time or he would take the matter to this mailing list.

But I, after thinking about the matters and seeing that I may not be
able to solve issues that I don't necessarily agree that are issues and
have been probably created because of possible shortcomings of my
actions, I decided to write to this mailing list myself.

I wish to apologize for the very long email, but this may be the only
way of ending this frustrating debate. I also apologize to Behdad, as a
dear friend and colleague, in case I have frustrated him or I am
continuing to do so. I consider all this to be for the good of the GNOME
Persian translation effort, and I believe Behdad considers the same.
While I deny forming any kind of conspiracy about GNOME translation, I
agree that I had made some mistakes, but I insist that they had been
because I had believed they would result in better Persian translations
in GNOME.

Looking forward for your constructive comments,
Roozbeh