On Mon, Sep 12, 2011 at 8:31 AM, Simon Phipps <si...@webmink.com> wrote: > [Recombining the thread] > > On 12 Sep 2011, at 12:43, Ross Gardler wrote: > >> On 12 September 2011 12:34, Simon Phipps <si...@webmink.com> wrote: >>> >>> On 12 Sep 2011, at 10:55, Ross Gardler wrote: >>> >>>> We need to manage this carefully. A Japanes language list to ensure >>>> non-English speaking people are able to participate in the project is >>>> fine. A Japanese language list for creating a different version of OOo >>>> for the Japanese market is not fine. >>> >>> The reality is likely to be somewhere in-between. For example, the PT-BR >>> localisation of OOo was the subject of extensive discussion in Portuguese >>> about exactly how to translate various aspects of the UI, none of which >>> would be of great relevance to English-speakers but which was still >>> development discussion. The same would be likely to apply to every locale. >>> >> >> Let me clarify "different version" I meant significantly different, >> not just a translation. > > You say "just a translation" but the debate on the PT-BR version led to two > competing releases for a time, with an impact on the community there which > lingers to this day. Localisation of a consumer application is never "just a > translation" as might happen to the strings in a server project; substantial > end-user decisions are debated, negotiated and agreed by thoughtful > developers. >
Correct, localization is more than just translation. But enabling the software for localization is a something the core developers need to be aware of as well. And the things that are done to enable localization for one language can impact the translations of other languages as well. So many of these things require (or at least would benefit from) coordination at a project level. > /The/ key reason for the success of OpenOffice.org is that there exists a > large, global community of groups of localisers who each act in autonomy or > semi-autonomy to create the release for each locale. Your message is a > wake-up call that we need to put a lot more thought into how the project will > approach them, especially if they will need to be separate projects in order > to retain their locale-specific autonomy. > It would be useful if you explained what you meant by "autonomy" so it is clear about what you are talking about. I see it as boiling down to two things: 1) Freedom of action and 2) Exclusive control Within an Apache project, committers have significant freedom of action. They can touch any part of the source code, any part of the website, any translation string in any language. With this freedom of action comes responsibility (don't break anything), and checks and balances (any other committer can veto your changes). But this does not come with exclusive control or "ownership". Your changes can be reviewed and vetoed by any committer, not just ones that you select, not just ones with the same native language as you, and not just ones working on the same component as you. And there are subtle and not-so-subtle social pressures to ensure that this veto is not abused by PPMC members who just want to comment on every proposal or change that comes by ;-) It might be useful, when you use the word autonomy, to remember that we're talking about individual committers (there are no corporate or organizational members of an Apache project -- we're all individuals) and their ability to contribute to the project. We're not talking about exclusive control of one group within the project over project resources. Of course there will naturally be focused areas of competency that we should respect. Not everyone here understands the implications of a 64-bit Windows port of Calc. But that doesn't mean that we create a ooo-calc-win64.incubator.apache.org group that makes decisions outside the review of the PPMC. Though naturally the members with that expertise, through their reputation, will have great influence in their areas of expertise. Ditto for Brazilian Portuguese localization. It is not control. It is about expertise and respect. -Rob > On 12 Sep 2011, at 12:44, Ross Gardler wrote: > >> On 12 September 2011 11:50, Ian Lynch <ianrly...@gmail.com> wrote: >>> If there is to be a NL build of the AOO product to be >>> released, presumably that build will take place at Apache? Or could it take >>> place elsewhere but only be formally released by Apache? >> >> It depends on what you mean by "takes place". Anyone can build >> anything they want, wherever they want. However a formal release of an >> Apache project must receive 3 binding +1's. The vote to get those >> votes *must* be carried out here on the official dev list (this one). > > So the release of (for example:) a new PT-BR binary needs three binding +1s > on this (English-speaking) list? > > S.
