Ubuntu's move into mobiles, tablets & replacing X.11 - is Mark Shuttleworth insane?

[liam_on_linux]

(Another repurposed mailing-list post. Please feel free to rip this apart.)

Maybe Ubuntu are not desperate. Do not forget Shuttleworth's wealth. He
can fund this for a long time.

Note also that a dot-com millionaire - not a billionaire, a "mere"
millionaire - is able to fund his own private freaking *space
programme* these days, with *multiple launch sites* mark you, out of
his own pocket without outside investment. While also running the most
famous pure-play electric-car company on the side.

Building a Linux OS by comparison is pocket-money stuff. The SABDFL is
worth two-thirds of a billion bucks and he lives in an offshore tax
haven. He can afford it.

It is a gamble, yes, but I think the gamble is this, laid out step-by-step:

[1] The phone market is huge and growing fast
[2] ... mostly at the cheaper end, at the expense of the
way-finder/ice-breaker Apple
[3] The tablet market is also booming; Apple still leads here, Android is weak
[4] The PC market is declining & desktop Ubuntu has not had a large
success here, just OK
[5] On phones, a Linux is the strongest player, but it's weak except on phones
[6] On tablets, MS is very weak with a divided offering
[7] On PCs, MS' offering is also weak & binary- *and* API-incompatible
with its tablet/phone offering
[8] Nobody, not even Apple, has a single offering across all 3 platforms.
[9] Canonical has a proven offering in the PC space & has been
showing, incrementally, that it can play in the phone, tablet & hybrid
space
[10] Also, in phone/tablet space, Canonical's offering leverages
(sorry) the fact that Android is also a Linux and Ubuntu can use
Android kernels & drivers
[11] Therefore the time is right to make a triple-play offering: a
single OS that can exploit all 3 and do it well, without major new R&D
investment

Secondly to all that:

[12] Android doesn't use X.11, it uses OpenGL
[13] X.11 has well-documented problems & weaknesses
[13a] Indeed, the only really successful desktop Unix doesn't use X.11
but uses its own display system, based (now) around OpenGL. It offers
an X.11 server as an add-on if you need it, though few do; but those
that do are happy.
[13b] The Linux desktop app market is actually getting moderately
mature & modern apps either talk to Gtk or Qt (or possibly SDL, OpenGL
or something); they all go through toolkits and libraries (.g. Cairo,
Clutter) and not direct to X.11. In fact today raw X.11 apps are
really pretty rare.
[13c] In the mobile space, X.11 has *no* presence *at all* & mobile
Unix apps all use indirect APIs and OpenGL.
[14] Modern GPUs scale down well, are integrated into CPU cores &
SoCs. OpenGL can be taken as a given now even on phones and tablets.
It is a very mature platform/API & vendor-neutral.
[15] The Linux kernel now understands graphics, has drivers and
mode-switching etc. X.11 is getting marginalised - X.org doesn't even
set its own screen mode by default -- graphical boot sequences do it.
[16] Therefore, the time is ripe to build a truly FOSS OpenGL-based
display stack for Linux that runs on mobiles (like Android) and on
desktops (like Mac OS X)

Yes, it's a big move. Yes, it's risky. But the time is right and I
suspect that their decision-making process was something like the
above.