On Wednesday 08 January 2014 17:15:04 Jos Vos wrote: > On Wed, Jan 08, 2014 at 09:44:50AM -0600, Lirodon wrote: > > Now, what I was thinking about was trying to see if we could unify as many > > of these EL spins as possible to produce a sort of "unified", independent > > distribution. [...] This is always the intention, but the outcome is nearly always the same as what happens with "unifying standards": you wind up with yet another standard to add to the stack and have complicated the very target of your reduction efforts. This is at the heart of the problem development firms have when trying to target Linux (and the idea we can rely on packagers to do things for us is not congruent with the way the consumer software market works, though it (mostly) works out for community interest projects and dire-need server-side software like Apache). It is important to note, I think, that the fear surrounding this has nothing to do with closed VS open source -- we're talking about open source. The real issue is unpaid community access to commercial project source in a form that doesn't require a massive repacking/development effort, and how readily that translates to user-end binary access in a form that is stable and maintainable. I'd say Red Hat has the best record of any company out there when it comes to this point, and though I heavily disagree with a (very) few things they've done, my disagreements have been technical in nature. If anyone has earned the benefit of the doubt, its Red Hat. (Arguing against fiat imposition of new standards or core components in Fedora, for example, really boils down to me being upset that the sponsor/hand-that-feeds is forcing a code change I don't like, regardless how my emotions make the argument feel at the time.) > It's pretty interesting to see how open source people (I'm one of them) > in a lot of cases tend to fight with each other instead of fight against > "the enemy" (TBD, but I'm talking about the non-open world, the "pseudo" > open source companies, etc.). We have a tendency to insulate ourselves so thoroughly from the *real* closed- source world that we forget how vast the darkness is and begin arguing amongst ourselves over what shade of white the light should be and which direction it should be pointing. "GPL" has evolved a social meaning quite apart from its legal definition, in particular there is a built-in expectation that any company that develops open source software is somehow mandated to provide free global access to source, a community space for people to argue over how evil the company is for not breaking itself trying to providing free binaries for every platform as well, and tacit or direct sponsorship for community projects that may not advance the company's goals (or may indeed even hamper them). For an application developer (as opposed to a platform developer), its actually easier to keep a clean, friendly reputation by writing purely closed- source software for purely closed-source platforms because the incidence of name-smearing will be near zero so long as they don't venture into the turbulent waters of the OSS community space. Adobe was considered a generally benign, and at times benevolent, community force until it released a platform (Flash), and ever since Adobe has been the target of untold trash talk. Taking the above in perspective, it is unreasonable to begin turning the spincrank on the conspiracy generator because Red Hat has decided to embrace (yet another) open source community. In fact, it is unhelpful to do so until there is a genuine cause to start FUDding the largest, most visible champion of open source software in the world. > Unless proven otherwise, I would give Red Hat the benefit of the doubt > (which they deserve, given their history). If CentOS provides an open > RHEL clone (more open than they did in the past), then we should accept > this and be (very) happy with it. Everyone is free to start their own > pet projects, but as long as there is no need for it (and again, given > the history of CentOS and their lack of transparency there was a need > for it in the past), any effort to start a "competitive" project is > IMHO a complete waste of effort. > > And about the reasons behind this: I do not know them, but maybe one > of the considerations is that an "official" clone is better than > having a company like Oracle gaining market share with their clone. This last bit is what I was really getting at. Time will tell, but this appears to be a masterful move to not only provide an obvious support plan choice that drives revenue right back home, but also undermine every other RHEL clone. Savvy, not evil.