Re: Complaint of the Slovak coordinator

[Jiri Eischmann <jiri eischmann centrum cz>]

______________________________________________________________
> Od: "Mattias Põldaru" <mahfiaz gmail com>
> Komu: helix84 centrum sk
> CC: gnome-i18n gnome org, gnome-sk-list gnome org
> Datum: 17.05.2010 02:35
> Předmět: Re: Complaint of the Slovak coordinator
>
>Ühel kenal päeval, E, 2010-05-17 kell 02:16, kirjutas helix84:
>> On Mon, May 17, 2010 at 02:04, Mattias Põldaru <mahfiaz gmail com> wrote:
>> > Ühel kenal päeval, P, 2010-05-16 kell 21:55, kirjutas helix84:
>> > Not to waste any resources (which voting one or the other out would do,
>> > not sure about Marcel, but Peter probably would leave because of that,
>> > just my impression), you could have two subteams, both coordinate their
>> > work to their best.
>> 
>> As I understand what you wrote, you may have missed that Peter was
>> already elected. This topic was raised because current coordinator
>> refused to step down.
>> 
>> > Set up local svn server and give any translator account and the right to
>> > commit there. When commit is done, the server sends emails with diff [1]
>> > to the team. Anyone can review and respond.
>> >
>> > Let's say Peter is the first warden of svn for he's subteam and when he
>> > responds to an e-mail of diff with "approved to be committed upstream"
>> > message, Marcel either commits it as it is without reading, or if he
>> > stumbles upon mistakes, he could ask these to be fixed.
>> 
>> Basically, nothing prevents this workflow currently as Peter has had
>> reviewer rights for some time. The problem is translations are stuck
>> in the "Committing" state because Marcel won't commit them without
>> going through them completely.
>> Even in the case you suggested the bottleneck remains. But this was
>> not the only issue with the current coordinator.
>
>No-no. The main idea was that he would give up the head reviewer role
>for some amount of modules, so whenever your subteam head says some of
>your modules is ready, it is to be committed or mistakes pointed out in
>reasonable time (whatever time you agree upon, three workdays?).
>
>What this also means, is that Peter would then be responsible for the
>quality of he's subteam modules. So I bet you will have hard times with
>all the fixing, because I have no doubt Peter will bring reviewing to a
>whole new level both speed and quality wise.
>
>Also you could have nice and friendly race for quality among teams :)

I don't see the point of doing this. I think the result is more than obvious. Marcel's part would have no or very little progress as it has had for several years and Peter's part would have dynamic progress, but some people will always question quality. And they would not have hard times with all the fixing. The Czech team has worked Peter's way for several years and we don't have a higher number of bug reports and the Czech translations are considered much better than the Slovak ones. For a user, it's much better to have 100-percent translated UI with 80-percent quality than 40-percent translated UI with >90-percent quality. The Slovak team will NEVER reach a higher percentage of translations the current way. One thing is the bottleneck caused by the fact that all translations have to be reviewed by the coordinator and the other thing is that the team will not keep translators because who would want to wait for several months until his translations are committed. Marcel woul
 d have to work on it fulltime to achieve a reasonably long time.

Users are not satisfied with the current situation, most translators as well, so why to choose a kind of half-solution you propose? And if your solution is chosen what will be metrics to decide which way will have turned out to be better? Will it be a rise in percentage of translations? I think no one here doubts that Peter's way will have better progress. Will it be a number of bug reports? It might be a sign that users have started using the Slovak translations again and work done in the bazaar way gets better over time mainly due to response from public. To decide what subset of the team works better is not that easy. At last but not at least most translators will be in Peter's team. As they have already mentioned 6 translators voted for Peter, 3 were for Marcel. But that doesn't mean they were satisfied with the current situation. They just didn't want to change a coordinator, but as far as I know some of them wanted fundamental changes such as quicker comitting, too. So 
 there is a pretty high change they would work for Peter's team as well. It would be for example Marcel + one or two translators against 7 or 8. Although it also shows a lot I don't think it would be a fair race. So if you propose a solution you have to mention its outcomes and the next step.

I suggest that we give Peter a chance. We can evaluate the overall situation in the team after some time, after a year or around GNOME 3. If it turns out not to work Marcel can always become a coordinator again stronger than ever before.

Jiri