I saw it coming. First Marc Fleury comments about Apachecon in his backyard, complaining about a missing invitation to attend, then Jim writes a somewhat resentful yet ironic entry, inviting Marc to follow the Apache Software Foundation business more closely, possibly putting some of his RedHat loot to good use by sponsoring Apache who, after all, has been the giant’s shoulder to stand on for a lot of JBoss business. It took just a few days and some Thanksgiving turkey for Marc to miss the afternoon football match and send another note to the blogosphere, poking more fun at Apache. The "fat ladies drinking tea" have now become a bunch of politicians hanging around as aging Star Trek fans, forgetting about code and preaching the Community gospel as if they owned the key to the Ultimate Open Source Truth.
One thing I’ll say for Marc: he’s damn good in rhetorics. A master of straw man, he depicts Apache as a group of oldies who only care about doing politics and duplicating stuff to indulge in their license talibanism. He’s not so good at fact-checking, though: a quick search at Ohloh would have shown how people like Geir Magnusson actually have bothered writing quite some code, other than helping freeing up the Java world within the JCP. Which kinda shows how Apache really is about writing code: we just go beyond that and care about the big picture as well, and we’re lucky to have people like Geir around, who enjoy being a pain in the neck to whoever tries to circumvent our ability to build and openly develop software for the common good.
It’s no wonder Marc is getting it all wrong when it comes to Apache being duplicating code almost for the sake of it (or, better, for the sake of the BSD license). Something Matt seems to resonate with. Unfortunately, they are both missing the point: Apache has been duplicating efforts for the sake of community based development, and BSD-family licenses just happen to be an enabler for such an endeavor, as they encourage participation from a diverse number of interested parties. It’s not that GPL couldn’t work at all, but let’s look at the facts: GPL as it stands nowadays is far from being a collaboration platform. Commercial Open Source is using it as a big poster sign of software built behind locked doors, with incidental help from the community but no real incentive to participate. Understandably so, to some extent: the moment software becomes a core corporate asset, the latest thing you want is someone on a different payroll to decide what your roadmap should be like. Which, in turns, brings you to MarcF camp of a world split in two: producers and consumers, suppliers and customers. To be sustainable, this old world has to rely on well-established economics, where customers pay for technologies funded and built by vendors. If that doesn’t happen, vendors have the right to whine and mope about: Open Source here is little more than marketing sugar on a granny pie.
Too bad this is as far as it can possibly get from the way Apache sees the world, which is about building a collaborative environment where diverse interests converge into building tools and software for a number of different purposes, from fun to profit (and possibly both). It all starts from people willing to scratch a common itch, understanding how joining efforts, exchanging ideas and building code together within a neutral environment allows to solve a technical problem faster and better. It enables different business purposes, as it doesn’t really matter if you want to build Open Source software for the sake of it, if you just want to use some stuff or if you want to run an international business around it. It mirrors the collaborative world we are living, where all lines are blurred and it’s difficult to say who is a customer, who is a vendor and even what on earth the product is. This is a world of shared R&D cost, which can leverage a number of innovative economics and doesn’t necessarily have to rely on licensesubscription sales to be sustainable.
So, is Apache duplicating code? Yes, indeed, sometimes even internally (just have a look at the number of feather-decorated web frameworks). Is that bad? Hell, no: if we were to think about distribution terms not being a reason good enough to justify duplication, quite a bit of Open Source would have to go anyway as it’s mostly about superseding proprietary alternatives. Is it suboptimal? Maybe, but the trade off between unencumbered access to the development process and mere source code availability, with right to fork as the only alternative, makes me welcome duplication. Does it always work? No, it would require a thick pair of pink glasses to claim that Apache projects have always been immune to corporate pressure and hidden agendas, yet in the long run there is little a corporate behemoth can do when confronted by peer pressure under technical and meritocratic terms.
All in all, I think I’ll stick to my Klingon lessons: Q’Apla, MarcF!