On Sun, Dec 18, 2011 at 9:26 AM, Graham Lauder <[email protected]> wrote: > On Friday 16 Dec 2011 12:03:53 Rob Weir wrote: <snip> >> >> Well, that corporate subsidy is gone. The question is how we can >> create a sustainable model for this project. What worked before, >> based on corporate charity, is not really relevant anymore. The >> question is, what works now? > > I have no problem with that, in fact I'm glad we're not beholden to a single > corporate entity. I'm also happy to admit that I'm glad it's at Apache. We > have a different model now, but that doesn't mean the other was a failure and > any talk that it was, is simply not backed by the evidence and > counterproductive in terms of community building. >
In other words you think my views are heretical. So here's the evidence that everyone can see. OOo, the project, had highly concentrated control in Sun. This was both direct organizational control -- their employees where heads of most of the engineering projects, and they governing bodies had reserved seats for leads, as well as concentrated investment by Sun -- most of the core developers were Sun employees. This dual concentration resulted in two failure modes that brought down the project. 1) the concentration of power and the refusal to devolve this into a foundation structure and 2) the non sustainability of Sun's investment. Both of these failure modes came together when a large portion of the OOo project broke away and formed LibreOffice and Oracle pulled out their investment. The fact that the LO fork happened and that Oracle withdrew is ample evidence of the failure of the OOo project. I don't know how else you can spin it. Of course, that doesn't mean that individual engineers failed personally, or that there were no good efforts. Far from it. You should never take such things personally. The nice thing about Apache, and one of the reasons I was an enthusiastic supporter of moving to Apache, is that it avoids both of these failure modes in the future. ASF is a long-established and strong foundation, with a governance based on individuals and meritocracy. Corporations do not join. We're not going to be able to recreate the power structures of OOo in Apache. This is a good thing. Also, Apache has protections against indirect control via concentrated investment from a single entity. We won't be permitted to graduate from a podling if the project is dependent on a single company, like IBM. That's the rule. So in order to graduate, we need to make progress technically with a product release, but also we need to diversify the investment base so avoid such indirect control. I think acknowledging the above is key to community building, especially in persuading LibreOffice volunteers that their ultimate goals are also fulfilled at Apache, and that Apache has far greater means, and proven ability, to prevent control from a single corporation than TDF does. -Rob
