[TriLUG] Steam coming to Linux

[Brandon Van Every]

On Thu, Nov 1, 2012 at 5:13 AM, Matt Hicks <mghicks at gmail.com> wrote:

> On Fri, Oct 26, 2012 at 6:13 PM, Brandon Van Every <bvanevery at gmail.com
> >wrote:
>
> > ... Valve did follow through on their plans.  They've supposed to be
> > doing a closed external beta of Steam on Linux sometime this month.
> > http://www.phoronix.com/scan.php?page=news_item&px=MTE5MzI
> >
> >
> The survey to enter the Linux beta is here:
> http://www.valvesoftware.com/linuxsurvey.php
>
> I took it. Seems rather focused on Ubuntu, but forum posts discuss that
> changing.
>

At the hope of starting a good old fashioned flame war :-), what other
distro would actually be sane for them as a starting point?  Last I looked
at various statistics, Ubuntu has a huge marketshare over and above the
other distros.  (Distrowatch is a joke; quote something else.)  They're the
only guys that focus primarily on usability *and* can claim extensive
testing of their package repositories.  I tried Fedora recently, and
they're doing the GPL Nazi thing about the NVIDIA drivers.  In Ubuntu I
installed them fairly painlessly using a few extra clicks; in Fedora I had
to chase down how to do everything myself, reading a bunch of gory
technical details.  Now I know the mission of Fedora is "bleeding edge open
source," and that the distro isn't really any good for maintaining long
term production systems, so that's to be somewhat expected.  The package
repositories were also really slow to load compared to the Apt setups I was
used to, unless you chased down yet more magic incantations to set up
repository mirrors.  It's pretty consumer unfriendly and if that's "the
best" RPM style distro, I question the relevance of RPM based content
distribution.

You can talk about the future of Wayland, Intel open source drivers, EGL,
and Gallium drivers all you want, but they're basically not here yet and
not ready for prime time. I have my doubts that they'll ever catch up to
NVIDIA proprietary performance anyways.  Even if they do, it'll be years in
the making.  Mesa is only just getting around to implementing the OpenGL
3.2 spec, and Khronos is out with OpenGL 4.3 now.  OpenGL is a
labyrinthine, terrible set of standards spanning 8 different recent
versions (3.0, 3.1, 3.2, 3.3, 4.0, 4.1, 4.2, 4.3) plus the ancient 2.x and
even occasionally 1.x stuff.  It drives implementors crazy, which is why
Mesa is so swamped trying to keep up and has only gotten as far as it has.
 Even Apple isn't faring much better, only being up to 3.3, but that's more
because they don't really care about 3D.  Half their revenue comes from
iPhones, so the latest greatest in 3D just isn't big money for them.

Supporting Linux native games "for real" would of course mean supporting
more than Ubuntu.  But given how fragmented and awful the Linux deployment
landscape is, it's pretty obvious where any sane developer would start.
 Consumer games are *not* about compiling source code to your heart's
content or diddling around with your favorite drivers for ideological
reasons.  These have been major impediments to a viable Linux commercial
gaming ecology and it's only now that logjam is finally starting to move.
 I hope that some other distros "get with the program" and adopt some
standards that make an indie game developer's life more sane.  Otherwise
we'll all end up beholden to Valve to provide the "pave over" solutions.  I
see Steam as good for a Linux gaming ecology, but frankly, that doesn't
mean I necessarily want to pay them to distribute my content.  Minecraft
never did it and they got filthy rich.  Also, what happens if Valve has one
"pave over" solution and Desura has a different one?   That sucks.

Because the logjam is moving, and Valve is applying some pressure, people
are at least now starting to think about updating some standards.  Like the
OpenGL ABI standard, which hasn't been updated for 12 years.
http://www.phoronix.com/scan.php?page=news_item&px=MTE4NDA

And before anyone starts shouting about GPLed open source games and how
that's "the answer," I have a substantial credit working on a really good
Battle for Wesnoth campaign, "To Lands Unknown."  Been there, done that.
 Being a Joint Author with a work that may have 5 to 20 other Joint Authors
is a commercial non-starter under US law.  AFAIAC RMS can stick it on this
particular issue.  Liberated Pixel Cup, bah, humbug!  It's like waiting in
line for toilet paper in the USSR.


Cheers,
Brandon Van Every