On Tue, Apr 29, 2014 at 1:47 PM, Marcus (OOo) <marcus.m...@wtnet.de> wrote: > Am 04/29/2014 09:47 AM, schrieb jan i: > >> On 29 April 2014 09:36, Jürgen Schmidt<jogischm...@gmail.com> wrote: >> >>> On 4/29/14 9:20 AM, Tal Daniel wrote: >>>> >>>> On Tue, Apr 29, Andrea Pescetti wrote: >>>> >>>>> I propose that, once a language reaches our release quality criteria >>>>> (currently: UI translation at 100% and maintained), we do not drop it >>>>> afterwards for the other minor releases. >>>>> >>>>> [...] I would remove unmaintained languages only when version 5.0 >>>>> comes. >>>>> >>>> >>>> Seems reasonable, to me, Andrea, I'm not sure that removing a language >>>> on >>>> major release should be so strict. What about removing a language only >>> >>> when >>>> >>>> a MINIMUM% of it isn't translated (e.g., 10%)? >>>> >>> >>> we had something like this before but defined a new rule to be 100% UI >>> complete and I think this is quite easy and a good rule. >>> >>> The case Andrea described above should be more theoretical if an active >>> community is behind a translation. We released Arabic with 3.4 but there >>> was no active community and nothing happened later on. >>> >>> I would still prefer the 100% rule. But anyway it's my personal opinion. >>> >> +1, not requiring 100% UI (which is quite easy to do for any translator) >> is >> a dangerous path. >> >> Nobody can today say when we do the next major release (5.x) meaning >> translations< 100% could be ongoing for a long period. For a minor >> release, its typically only a handful of messages that are changed, so it >> not a big workload for any individual. > > > However, from the view point of a normal user who just wants to update to > the next version, it would be confusing why no localized install file is > available anymore. > > So, from my side a clear +1 to keep these languages. > > How much we allow to be under 100% is just a question of definition (and > agreement). ;-) >
We want quality releases. % translation is part of quality, of course. But there are other aspects as well. Certainly looking at % completeness is easy for to measure, but it is not necessarily the best criterion. We want to avoid a situation where a translation is rushed and done poorly, in order to meet an arbitrary % goal. I'd rather have a high quality 95% than a low quality 100%. Of course, PMC members do not know all languages. So we need to rely on the translators and the local community. Maybe we can make a criterion from that? For example: If a translation is more than X% complete, AND if that language was downloaded in the beta release more than Y times, AND the RC was reviewed by the translator and Z other community members to vouch for having usable level of quality, then we include it in a release. Or some other way of having the local community take ownership of making this decision. -Rob > Marcus > > > > --------------------------------------------------------------------- > To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org > For additional commands, e-mail: dev-h...@openoffice.apache.org > --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org For additional commands, e-mail: dev-h...@openoffice.apache.org
