On Tue, Jul 22, 2014 at 10:49:33PM -0400, Bill Nottingham wrote: > Matthew Miller (mattdm@xxxxxxxxxxxxxxxxx) said: > > * Open Floor (mattdm, 17:17:32) > > * Translators appear to be moving over to Zanata now (gholms, > > 17:19:16) > > I'll bite. > > 17:24:08 <mattdm> Dimitris Glezos from Transifex makes a request for the > board to define a cost/benefit analysis of the risk of using proprietary > software like transifex in support roles > 17:24:48 <mattdm> No one has brought that to us formally (the transifex > thing, obviously, not beer)... does anyone want to pursue it? > 17:25:32 <mattdm> Dimitris' argument basically seems to be that the L10n > project should push to use whatever tools get the job done, and it's the > board's job to push back if they want to use something that is dangerous or > counter to ideals > 17:25:37 <inode0> This isn't something the board needs to decide if the > translation folks sort it out as they are doing really. > 17:25:51 <mattdm> inode0 yes, that seems to be my thought, exactly. > > To be clear; I'm fine with the board staying out of this if they feel > translation has it under control. But I can't help but think that how this > change happened was strange. > > Transifex has been effectively proprietary for almost two years [1]. It was > confirmed as such 16 months ago [2], and I recall being in some discussions > about it around or after that timeframe, although I don't have a link. > > So... why was this not an issue for translation until now? Were they merely > not aware? If the defining thing that made it not fit-for-purpose was the > hosted service not being F/OSS (as seems to be the case), is there some > effort that should be taken forth to monitor such things? > > And if this lack of open-sourceness is a bright-line issue for > Fedora-related activities, why draw the line here, and no further? Zanata > is, after all, hosted on github. While it would certainly be > self-immolating to refuse any random upstream piece of software that may be > developed or hosted in ways that are not considered fully 'free', it's > not as if Zanata has any development community outside of Fedora's chief > sponsor. Or, more closely - Fedora infrastructure uses github as well. > Fedora design has had flickr and deviantart groups. We've had Google > hangouts at FUDCons. We *shut down* Fedora Talk, rather than expand it. > > To look at it in terms of our foundations, re: building open source software > communities: > > The community includes current and potential or future contributors. Our > outreach begins with our free distribution, and we constantly develop ways > to give collaborators additional on-ramps for participation. > > Do as much of the development work as possible staying close to upstream > projects. We promote upstream communities by collaborating on patches, > providing the latest upstream versions for our development and testing > branches wherever possible, and making sure upstream products work > consistently and well in our stable releases. > > Do we collaborate better by using the services where our upstreams may lie, > regardless of the 'freedom level' of their implementation status, or do we > collaborate better by forcing them to come into our playground on our terms, > such as by tying translations to Fedora accounts and Fedora infrastructure. > > Now, if we're willing to say that our commitment to open-source software > trumps our commitment to how we build and expand communities and > collaboration, then yes, Transifex would fall short. But then I'd say > github for directly Fedora-associated projects would fall short too. On one hand, if the L10n team decides that question in favor of freedom, it seems odd to me that the Board would seek to back that decision out, unless migrating Zanata put an unreasonable cost on other teams. On the other hand, though, it seems to me we're talking about sacrificing the potentially higher community involvement that we get from a service like github (which seems to have all the eyeballs of open source aficionados). > But I'm not sure that's right - if we intend to foster communities and grow > collaboration in all areas, we need to go where those communities are. That > might mean translating in a wider community at Transifex. That might mean > collaborating with other developers on github (or bitbucket, or...). That > might mean having development discussions via Hangouts, posting the > resulting videos to YouTube. It might mean broadcasting our announcements > on Twitter. It might mean advertising events on meetup.com, and presenting > at other events hosted there. Related: To what extent we can query, capture or use information from these sources (github, Transifex) to discover and take advantage of community building opportunities? Is it possible to e.g. give badges based on work there, using some sort of connector/proxy/etc.? (Asked the non-webdevmonkey.) > Again - I don't want to change whatever decisions the translators have made. > But I wonder if doing this sudden switch based solely on the year-late > discovery of Transifex's status sets some sort of precedent (that we don't follow > elsewhere) where we sacrifice our goals of building broader communities in > favor of an 'only F/OSS' stance. Also important is to what extent this decision was driven by problems within the translation teams, and how those problems happened. Cited over the last couple years were issues with translating documentation. Of course, the tools for building documentation are driven by folks in Red Hat's documentation teams. So their presumably better integration with Zanata, which translators have been using for some (most? all) Fedora docs, is no surprise. If this decision were being driven over the objections of active translators, I'd be really uneasy about the motivations (and I suspect others would, too). It's hard to argue, though, when the discussion on the trans@ list seems to be mainly supportive of migrating to Zanata, although it's not a massive discussion. I don't know whether active translators are active on the trans@ list, vs. geo-community specific lists. It does trouble me that Transifex is probably the one thing out there in the larger world we in Fedora can point to and say: "That's a success story for Fedora, in the project's role as an incubator for ideas that *don't* have to come from our sponsor." So moving away from it, while it probably has little effect on Transifex's success, makes Fedora seem somewhat petty (and possibly hypocritical as may be implied above). > (As to Dmitris's suggestion of a cost/benefit analysis, I would find it naive > to think the Fedora board can cause Fedora to switch to something other than > the platform written and maintained by Red Hat i18n engineering and used by > RH L10N for other projects; the cost/benefit analysis that would drive such > a discussion would not be factoring the cost to non-RH translators in general.) On the other hand, it's not clear to me that every project internal to Fedora is being proposed to move to Zanata as a requirement. There certainly doesn't seem to be a lot of win in it for developers whose projects are being translated well in Transifex. However, if it turns out that those translations are largely being done by teams and individuals that would rather be using Zanata, then developers could be in for a surprise, and that wouldn't be good. -- Paul W. Frields http://paul.frields.org/ gpg fingerprint: 3DA6 A0AC 6D58 FEC4 0233 5906 ACDB C937 BD11 3717 http://redhat.com/ - - - - http://pfrields.fedorapeople.org/ The open source story continues to grow: http://opensource.com _______________________________________________ board-discuss mailing list board-discuss@xxxxxxxxxxxxxxxxxxxxxxx https://admin.fedoraproject.org/mailman/listinfo/board-discuss
