Software is [fF]ree just as long as nobody cares about it

While the title might sound inflammatory, I don’t think there is any way around this: there is not going to be any software that is entirely Free of every constrain unless nobody but its author cares about it. The moment enough people care about a particular piece of software, you have to apply constrain to that, which are usually in form a license. Sure the license could be a Free Software License, or an Open Source License, or both of them at the same time, but the very fact that you have to apply a license over it is usually enough to piss off enough people into saying that your software isn’t really free.

I’m one of those people who really can’t be told that there will ever be an “year of the Linux desktop”; I think that one of the reasons for that is that each year we witness the same amount of controversy in Free Software; enough controversy that even your average daily political issue wouldn’t snob it.

It was an year that I wrote about Sony dropping Linux support and we still are seeing people – declaring themselves Free Software advocates – defending the cracker and the “right” of people to simply copy games that weren’t designed to be free in any way to begin with. I have quite a personal grudge against these people for the simple reason that if they really cared about Linux on the PS3 they would have contributed to the work, rather than hiding behind that excuse just to be able not to pay the authors of content that was designed to be paid for. By accepting this position, the Free Software community is actually showing itself quite immature: “You should follow my (Free) license, but I won’t follow yours!”.

But if last year we had the “defection” of Sony and Oracle, this year started with a pissing match between Canonical and RedHat followers and the rest of the “community”. And before somebody asks, I have personal dislike of the way most of the Canonical deals go on, and I happen to agree with Jürgen about Jono’s average task about moving the spotlight away. But while I have a personal preference for RedHat methods – and a number of colleagues who I don’t hesitate to consider more than simply work acquaintance in their ranks and file – that doesn’t stop me to think that the kernel patch move is quite obnoxious for the community at large.

And now we’re getting a number of mixed signals about Google’s Android, including the mistaken controversy about Linux syscall interfaces — I can’t blame that on bad intentions though, as I was also confused by unclear license information and that is indeed a problem that should be solved. While I now can tell I did read about it before, I totally forgot about the syscall exception, and that should probably be better known.

There are obviously different ranges of “non-as-free-as-you’d-wish” to be found in software that people care about: from license restrictions to the way upstream and distributions collaborates… I have encountered an example of that just last week, with Quagga a routing daemon software.

I have taken up maintaining Quagga simply because I wanted to use it on my home setup, as it was the “easiest” way to get the ADSL router-in-PCI-card to know where to send packets coming from the IPoA bridge — note if you happen to have a Linux-compatible ADSL/ATM modem picking up dust I wouldn’t mind replacing that one… and if I did I would be helping picking up the linux-atm stack that seems to be bitrotting in Gentoo.

When I picked up Quagga I did some polishing among the other things because it was originally copyrighted by a non-Gentoo-related company who contributed ebuilds and init scripts, and their copyright statements took over the Gentoo ones up to that point; that was probably a signal I failed to interpret. While I tried sending upstream my patches, the upstream mailing list has shown a very low level of actual involvement from the one developer and can probably explain why two features that the ebuild did support are still external patches.

At any rate, the 0.99.18 version that was recently released turned out to be not so good as it was intended to be, and the homepage now shows a pretty interesting warning that you would need three patches to fix a few issues that came up after the release — which has been reported to Gentoo, and were fixed before the 0.99.18 ebuild hit the stable tree, for what concerns Gentoo users. When it was proposed on the mailing list to release a bugfix tarball, I proposed simply getting the distributors together and manage a shared “blessed” list of patches and backports for Quagga, to release the same way Greg releases the stable kernels. I’m not really convinced by the reply from the one developer above:

If someone wants to maintain a stable series, they’d be welcome. TBF
though, this is what some distros already specialise in. Some will
even answer your phone calls and try debug & fix your problems for
you!

(FWIW, I’d strongly encourage anyone using Quagga in a business
setting to get a support contract with their favoured vendor; if
they’re a general Linux/Unix vendor rather than a network/routing
specialist, then be sure to tell them Quagga is an important package
to you. This is what helps pay for people to work on Quagga.).

Maybe I’m reading too much between the lines, but while I do understand the need for people to be actually paid to work on stuff (I definitely don’t subscribe to RMS’s doctrine that developers should be okay with making “a mere living” — we’re the bone structure of this age!) I don’t see how inviting vendors to forget coordination and collaboration to try spilling money from users.

Are we really at this point? If we are, that’s a sorry state for Free Software to be in, and is likely going to be a signal of a crumbling system.