It already is. It already has been. It will continue for the
My scorn for lazy tech reporters who have made an industry out
of "This is the year of Linux!/ No it isn't, stupid hippies!" is no
secret. In fact I could make my own little industry out of scorning
lazy, useless, content-free tech reporting, except that complaining
all the time is dull and annoying. It gripes me greatly that this
grand bully pulpit of online publishing, where a single person has
the potential to reach the entire world, is wasted on 90% dreck.
But then Theodore Sturgeon said that ninety percent of everything
is crap, so I guess that's just the way humanity operates.
So, getting back to The Year of Linux. Those of us (like you, my
fine readers!) who are actually clueful about the tech industry
understand that there are very few Aha! events, which are dramatic
events that change forever life as we know it. That's a Hollywood
fiction. (Maybe that's why so much tech journalism is so awful--
too much TV.) In fact I can't think of a single genuine Aha! event
in my lifetime. Putting astronauts on the moon was the culmination
of decades of hard work. After more than a century of automobiles,
the primary advance is more cupholders. Our modern personal
computers that give us mighty processing and storage powers for
small price tags are the results of years of refinements and
incremental advances. Fast advances to be sure, but not Aha!.
Linux is a relative infant at age 18, so Tux isn't even legal
drinking age yet. Going from zero to where Linux is today in only
18 years is remarkable, especially when you compare it to poor old
Windows. The first release of Windows was in 1983. You'd think that
with 26 years of development and tens of billions of dollars it
wouldn't be quite so awful, but maybe that's an indicator that
Sturgeon's famous quotation is more than a wise saying, it's a law
of nature. But again, no Aha! moments.
What Does Winning Look Like?
We like to joke about world domination, but I would settle for a
level playing field, a computer marketplace that is not controlled
by an evil, rapacious monopolist. I would settle for hardware
vendors that support Linux equally, and for gosh sakes what could
be easier? Release the specs, donate some machines for testing, and
let FOSS developers do the rest. I would settle for retail outlets
that are honest about customer demand and actually meeting it,
instead of letting Redmond tell them how to run their businesses,
and then making lame excuses. I would settle for legislatures and
parliaments and standards bodies that do their jobs, which at least
in theory are not propping up Microsoft.
Winning looks like honest reporting and product reviews written
by people who have actual knowledge and do actual research. In my
dreams the giant thundering hordes of astroturfers would all be
smitten with attacks of conscience and go mute. I know this breaks
with established industry practices, but I can dream.
When you think about it, for all their size and clout, Microsoft
sure needs a lot of outside help to stay on top.
The only market segment that Linux does not have significant
penetration is the desktop, though even this is debatable given the
difficulty of coming up with accurate figures, and the well-funded
opposition to coming up with accurate figures. In all other arenas
it is a force to be reckoned with: supercomputing, datacenter,
embedded, networking, servers of all kinds. This is no small
achievement for an OS that has had to succeed strictly on merit and
against relentless opposition.
We need to give Linux a lot of credit for improving rapidly on
multiple fronts all at the same time. Complaints get all the
headlines: wireless sucks, KDE4 sucks, video sucks, support for
Microsoft document formats and protocols sucks, etc. Yeah whatever.
Since the 2.6 kernel was launched progress has been phenomenal.
Unglamorous things like shiny new PATA, SATA, SCSI, and USB.
User-configurable I/O scheduling. SELinux and AppArmor, overhauls
of key subsytems like Upstart replacing init, udev/HAL, and
substantial progress in wireless and video. The Linux wireless team
have done an exceptional job of rewriting the entire wireless
subsystem, and bringing hardware vendors on board.
At the same time usability and polish have been getting a lot of
attention; and different packaging like LiveCDs, bootable USB
distros (which Windows still does not have except via some
third-party tools that don't work very well and require the full
retail Windows version), PXE boot, wake-on-LAN and network
installs, great tools for easily making your own "re-spins", nice
growth in application software of all kinds. Powerful testimony to
the power of free software. In contrast, after all these years of
supporting Nvidia what do we have? The same old closed binary
drivers. At least with Intel and AMD/ATI we have hopes of attaining
the same status we had in the late 1990s, which was several
different 3D video chipsets with FOSS drivers.
The Last Link: OEM Desktop Linux
As so many folks have been saying for so many years, the key to
visibility is the OEM channel. We should be able to go to any
computer store, whether online or in meatspace, and see
good-quality fairly-priced Linux machines sitting side-by-side with
Mac and Windows. I know this strikes some people as in insane
concept, which means they don't even notice there are multiple
brands of PCs, servers, networking gear, hard drives, monitors, and
other components all sitting quietly together.
You might have noticed that Apple could never get good shelf
space, but was always relegated to its own ghetto. CompUSA had an
Apple section stuck way in the back of their stores, and Apple had
to open their own stores to get anywhere. Even Fry's shunts Apple
to a dank corner.
Good OEM Linuxes mean that Linux itself doesn't need to be
perfect in every last detail, because then it's up to the OEM
vendor to release a machine where everything works the first time.
We hear the same stupid FUD all the time: oh noes, too many
Linuxes!! Hardware support is too weirdz!! Give me a break, nobody
can be that dumb without getting paid for it. The OEM vendor picks
a limited, well-supported set of hardware; contracts with a Linux
distributor like Canonical, Mandriva, or Red Hat for support and
engineering, and voila! Instant excellent Linux box. How hard is
that? Apparently the concept itself is monumentally difficult to
grasp if we go by the awful netbook Linux customizations the titans
of industry have been releasing, or high-ranking industry analysts
Lenovo's Matt Kohut. On the other hand it's obvious, logical,
and not all that difficult if we go by the examples set by
independent Linux vendors like ZaReason, System76, and Penguin
Computing. And any random moderately-accomplished Linux geek.