FREE SOFTWARE IS BUSINESS (RE: ZLG)
ZLG recently posted a long piece about the politics of free
software and why it shouldn't/can't be ignored:
gopher://zlg.space:70/0/articles/Software/./free-software-is-politics.txt
gopher://zaibatsu.circumlunar.space:70/0/%7ezlg/0036_free-software-is-politics.txt
It starts around the topic of Systemd adoption being debated in the
Alpine Linux community. The endless Systemd debate is definitely
political, I agree (and I'm on ZLG's side, against it, as well),
but I think the interesting thing is the force behind it.
Here I'm not talking specifically about Red Hat, but in general
about commercial interests in the Linux ecosystem. In short you've
got the divergents, key examples being Apple and Google, who took
free UNIX code as the basis for their own Operating Systems but
developed them independently from that point. Then you've got the
collaborators, in particular Red Hat and SuSE, who made Linux
itself their product, taking market share away from the old,
proprietary, commercial UNIX systems.
The divergents need to maintain a complete software environment for
their derrived OSs, funded by hardware sales (of hardware which, as
it happens, usually doesn't have very good support in Linux). On
the other hand the collaborators have a lot of work done for them
by all the different companies and individuals contributing to both
the Linux kernel itself, and the huge variety of software running
atop it. With things like X, they've even been using code
originally written in large part by the old commercial UNIX
companies that they (to some extent) replaced.
The trouble is that they want to build the best product to suit
their particular clients, who don't represent Linux users as a
whole. So if they decide that their clients are best (or most
cheaply, in terms of development costs) served by changing a basic
feature of Linux distros such as the init or windowing system, the
only way they can pull this off while retaining the advantage of
having other people write and debug code that works in combination
with it, is to push for its widespread adoption. If they don't
achieve this, then they'll just end up with their own equivalent of
Apple's Darwin or Google's Android, started from other people's
code but relying entirely on their own programmers to move it
forward down a separate path.
So maybe for users things like Systemd adoption are political
issues, but for a Red Hat software project team it's a matter of
existance. If they develop something like Systemd and _can't_ get
it into the wider Linux ecosystem, then it doesn't get integrated
into the collaberative environment that they're relying on. Other
people don't design their software to integrate with it, don't
write tutorials for it, and generally leave it up to Red Hat to
make everything work themselves. Red Hat don't have anything like
the money of Google or Apple, so in the long run they can't afford
to let that sort of divergence take root.
The benefits from commercial collaborators in the Linux ecosystem
flow both ways of course. With X for example, Red Hat filled a gap
from the big corporations such as Sun, DEC, and IBM, who once
contributed to the open-source reference X implementaion from which
their proprietary X servers (and the free XFree86) were derrived,
in order to maintain compatibility. Now that the Red Hat developers
have all moved over to Wayland, the Xorg server's future currently
relies on a release manager working in his spare time for Patreon
donations:
https://www.patreon.com/p12tic
In this regard the opinions of users, and even individual
developers, are easily overwhelmed by the basic power of a company
paying developers to work on stuff. Whatever politics goes on in
the mailing lists and forums, the commodity of open-source software
is code, and while collaborating, people are also selling their
code in exchange for influence. As well as a political environment
it is a market environment, producing for some a free device to
serve a function they need themselves, and for others a product
they can sell on for profit (even by the indirect means of
commercial support).
So far this hasn't worked out all that badly. I don't like Systemd
or PulseAudio but I can generally avoid them. Though edge cases do
creep in such as setting up a VPS with a service that only offers
Systemd-based distro images for use, and Firefox offering better
support for PulseAudio sound than ALSA. But free open-source
software is still young as a concept, and it's not clear whether
the balance of power (which here is the balance of code) will
remain steady, so that the options for users remain open.
- The Free Thinker.