Annoucement: New translation status page is installed

Bernd Groh bgroh at redhat.com
Tue Jul 6 07:51:17 UTC 2004


Christian,

just so you don't think I'd intentionally not answer this response of 
yours, to avoid an issue or for whatever reason.

Christian Rose schrieb:

>I realized I haven't answered this. As there was a comment about my
>supposed attitude towards new contributors, I probably should.
>
>ons 2004-06-23 klockan 02.55 skrev Bernd Groh:
>  
>
>>>I'm also confused by the "Translator" field. For the anaconda sv.po,
>>>it's filled in with a name I've never heard of before. Even worse, this
>>>person's e-mail address is a Danish one, and the domain part is a very
>>>dirty word in Swedish...
>>>      
>>>
>>That would be the person who currently has anaconda assigned.
>>    
>>
>
>Ok.
>
>
>  
>
>>>Since 2000, I've been working from scratch on doing translations for Red
>>>Hat and now Fedora, with the goal of keeping them high quality and 100%
>>>for every release. We're currently two persons doing this work, and
>>>we've managed to do exactly this for a *lot* of releases by this time.
>>>We've recieved quite a lot of positive feedback about the quality, too.
>>>
>>>Thus it's not really exciting to see that any random bozo can suddenly
>>>take over control over a Swedish translation and fill in dirty words.
>>>I'm not amused.
>>>      
>>>
>>Do you want me to not give someone access because s/he's got a danish 
>>email address and the domain name is a dirty word? Regardless of how 
>>fluent this person may be in swedish, and how much this person would 
>>like to participate? Do you know for a fact that this persons 
>>translations are bad? If so, let me know, and I disable this persons access.
>>    
>>
>
>Let me make some things clear:
>
>* I have nothing against Danish people, in fact, I live close to
>Denmark, I often visit Denmark, I regularily travel to GUADEC:s with the
>Danish contingent, and stay where they are staying, and enjoy beers and
>their company. Some of my best friends are Danish.
>That being said, when someone with a Danish mail address starts
>translating into Swedish, I raise an eyebrow. There can of course be
>many plausible explanations, but it's definately not what you'd expect
>at first.
>
>* Of course anyone is free to use a dirty word in their mail address.
>But when that dirty word is a Swedish one, the mail address is a Danish
>one, and the person is supposed to translate into Swedish, one can start
>to wonder if this guy is for real, and not just some bad joke someone
>felt like making when discovering a form field open for writing on a web
>page.
>
>* This is the first time I ever hear about this person. Swedish is not a
>big language, and free software translations for it is of course done by
>an even smaller community. I've been translating free software into
>Swedish for six years now, and am quite familiar with the few people
>that are working to do the same in other projects.
>Searching Google, I can find no references to a person with this name
>and mail address. Searching only for the name leaves me with quite a few
>hits, since the first name and last name are both common ones. But no
>match whatsoever has a connection to Linux or OSS/free software or
>translations.
>
>Given all of this, I find it hard to believe that this person is for
>real, and I find it quite insulting that you believe that this person
>would do a better job translating anaconda than me. Perhaps that's not
>what you really and honestly do believe, but then again your comment
>together with the fact that this person has taken the anaconda
>translation, and thus in fact is allowed to lock me out, would sure let
>one believe this.
>

Maybe the fact that I've offered you the maintainership of all swedish 
modules, twice even, and now for the third time, since it's still yours 
for the taking if you want it, which would put you into the position to 
release the module from that person and take it yourself, or let someone 
else take it, should, IMO, not make you believe this. Or have you seen 
me offering the maintainership of the swedish modules to anyone you've 
never heard of? And no, I do not believe that this person would do a 
better job translating anaconda than you, fact is, I do not speak 
swedish and I cannot and will not make judgement. I know through 
third-party feedback that you've done an extremely good job keeping the 
swedish translations up to date and to a good quality in the past, and 
that exactly is why I've offered you the maintainership of the swedish 
modules, and we all would be happy if you'd do it. My statement didn't 
intend to say that the person who took anaconda has a right to look you 
out, I simply meant to say that you're invited to become a maintainer 
and release that person yourself, since I don't think it's my job to 
decide on who's allowed to do swedish translations and who isn't. I 
believe swedish speaking people should do that. I'm sorry if it felt 
like an insult to you, it wasn't meant to be one.

>>I understand what you're saying, but I cannot at all agree with it in a 
>>general sense. If you wish, I can make you the maintainer of every 
>>swedish file, then you will even be notified if anyone commits a swedish 
>>file. And while in a few languages it may be reasonable to not just give 
>>anyone cvs access, in a lot it isn't, since there is noone who could 
>>judge who should have access and who shouldn't. I personally prefer to 
>>trust anyone to do the right thing by default, and if they don't, well, 
>>let me know.
>>    
>>
>
>As you've probably guessed, I, like most other maintainers, don't trust
>people to do the right by default. I want a mandatory intermediate step
>(review) to catch potential problems *before* they get committed. This
>is nothing really new, ask any software maintainer... It seems absurd
>that we are even arguing about this.
>

As said, if there is a coordinator willing to make herself/himself 
accountable for this decision and the language group, in general, is in 
favour of that, I completely agree. If the language group prefers to 
have cvs access open, or there isn't an established group, I do not 
believe an intermediate step to be feasible. Who'd be to judge when cvs 
access should be granted? Now, I believe you are referring to the case 
where there is a coordinator willing to be accountable and the group 
being in favour of restricting access, right? If this is the case, I do 
agree with you, and I don't argue that point. If no, and your argument 
is a general one, then who should judge when cvs access is to be granted 
if there is nobody or nobody who wants to be responsible for such 
decision? Who is going to make such decision?

>Let me give an example of how I've learned the hard way not to trust
>people doing the right thing, even though it may not really be
>intentional at all. I don't believe people are automatically evil. I
>just believe people can make mistakes.
>
>Lots of free software use translation projects to organize translations.
>Some software maintainers think they can handle this organization
>perfectly well themselves though. Such an example is XMMS. I translated
>XMMS, but XMMS used to release not particularily often. I spend much
>time translating other projects aswell, and I don't always constantly
>monitor all the projects unless I know there's a release coming, and
>when of course it's then time to update.
>
>I had made the Swedish XMMS translation complete, sent it out for review
>by other translators several times, and incorporated the changes and
>suggestions. The result was a polished translation which I got several
>positive remarks for (which in case of a small language like Swedish is
>a lot).
>
>Suddenly one day there was a new version of XMMS that had been released.
>I wasn't aware that this release was coming. The XMMS developers had
>forgotten to inform translators. Worse, the Swedish translation had a
>lot of strange and sometimes amateurish translations in it. To make
>things worse, some people publically made fun of this, and as they knew
>I had been responsible for the Swedish XMMS translation in the past,
>pestered me about it.
>
>It turned out that some complete unknown guy had noticed that the
>Swedish translation of the development version wasn't complete at some
>time mid-cycle, completed it himself, and sent it to the software
>maintainers, who in the spirit of "ok, this looks more complete"
>committed it without asking any questions or informing anyone else, like
>perhaps me. Worse, it would take a whole year or so until they planned
>to do a new release, and this translation would be there for this whole
>time to make everyone either disgust it, or make fun of it. Certainly
>rather few people would enjoy it.
>
>This guy who completed the translation obviously wasn't at fault. He had
>good intentions, although his written language skills happened to be
>bad, and he just did whatever one would expect him to do and sent his
>work upstream.
>What was done wrong was at the software maintainer level -- since they
>didn't use a translation project structure with responsible teams and
>translators, but wanted to coordinate this themselves, they should have
>done exactly that. But they didn't, and just committed it.
>
>
>What scares me the most is that this is bound to happen again and again,
>not just with XMMS, but this time with the whole of Fedora, because the
>current structure by default allows anyone to commit whatever they want
>without even talking to anyone else first, before it's too late. A
>recipe for disaster.
>

In the new system, only if there is nobody maintaining the file. If 
there is a maintainer for a module, the maintainer is informed when 
someone takes the module, when someone does a commit. Even more, when 
the file is finished, the maintainer has to approve the module before 
others can take it again and commit changes to it. What is the major 
drama about having the changes in the cvs? If the translation is no 
good, you can always roll back. I prefer to have the changes in cvs, at 
least then everyone can have a look at the most recent changes without 
having to talk to the person who has the po-file in her/his inbox. A 
commit is not a final thing, it can easily be undone. Or do you always 
make sure that your most recent software changes work before you commit 
them? How do you do that with a larger addition? Do you wait till it's 
finished and not keep track of changes for days, but commit a whole 
bunch of changes at once? I prefer to commit whenever I do a minor 
change, or constantly commit even if I do a major change, even if things 
aren't working yet. And if I really should commit something that wasn't 
good, I simply get the previous version and keep working on that. Why 
would that be so wrong in our case? Or did I misunderstand what you were 
saying?

Now, why is the new system compared to the previous system an even 
bigger recipe for disaster? Doesn't it actually help avoiding the 
problems you were just talking about?

Regards,
Bernd


>Christian
>
>
>--
>Fedora-trans-list mailing list
>Fedora-trans-list at redhat.com
>http://www.redhat.com/mailman/listinfo/fedora-trans-list
>  
>


-- 
Dr. Bernd R. Groh                       Phone : +61 7 3514 8114
Software Engineer (Localization)        Fax   : +61 7 3514 8199
Red Hat Asia-Pacific                    Mobile: +61 403 851 269
Disclaimer: http://apac.redhat.com/disclaimer/






More information about the Fedora-trans-list mailing list