Red Hat: no desktop products coming [LWN.net]

Ubuntu/Canonical not non-profit

Posted Apr 17, 2008 15:18 UTC (Thu) by pr1268 (subscriber, #24648) [Link]

Well, then that just adds fuel to my argument that it's not impossible to make a successful (profitable) business out of consumer desktop Linux.

Thank you for the correction.

except that Ubuntu/Canonical has not actually made a profit.

Posted Apr 17, 2008 15:31 UTC (Thu) by pizza (subscriber, #46) [Link]

I don't know how much money they're losing, or even if they're losing more or less than they
used to, but Ubuntu is still one of those "charities" that RedHat alluded to.

except that Ubuntu/Canonical has not actually made a profit.

Posted Apr 18, 2008 8:24 UTC (Fri) by AlexHudson (subscriber, #41828) [Link]

I actually though that charity comment was aimed squarely at OLPC, with whom Red Hat did all
that Sugar GUI work.

The elephant in the room does seem to be Canonical in many ways, though. Even though Red Hat
don't ship a consumer desktop type product, they put an awful lot into desktop-related
technologies - that list in the article is pretty extensive really, and we know Novell do an
awful lot of work too. 

When you think of what Canonical contributes, it seems to be basically be bzr and upstart, and
various bugfixes. I can't think of much else.

except that Ubuntu/Canonical has not actually made a profit.

Posted Apr 19, 2008 2:14 UTC (Sat) by mmarq (guest, #2332) [Link]

Funny thing, and if i remember correctly, is that the whole Linux movement started by
"charitable" work on a i386... that couldn't be considered a server system by almost all of
measures!

But i think that this desktop vs server is getting to much bi-polarized and is full of "fuzzi"
rhetoric, specially for the possible Linux business models.

If RH ever had thoughted about making fortune selling "boxed" software... well, even MSFT can
do it less and less, to the point of rumor of selling online software "features"... but all of
this doesn't invalidate that the Desktop business is a no go... "au contraire"...

Considering autovectorization & autoparalelization

http://www.edn.com/index.asp?layout=article&articleid...

all other optimizations possible

http://llvm.org/pubs/2004-01-30-CGO-LLVM.html

Specially in the graphic arena, to the point of making graphic developers drool!...
http://zrusin.blogspot.com/2008/02/openvg-and-acceleratin...
http://zrusin.blogspot.com/2008/02/gpgpu.html

Well its pretty obvious what i want. I want autovectorization & autoparalelization with a
compiler framework designed to support transparent, lifelong program analysis and
transformation for arbitrary programs, with compile-time, link-time, run-time and "offline"
optimizations, specially for my desire for highly graphic workloads, to the point that i can
use one GPU or several, or better said GPGPU, of which every model of AMD ATI above the R600
is already a GPGPU because it can run the open sourced "firestream" SDK, and so offload the
CPU and get improvements that by Amdahl law i could get near 50% !...

Well i'm not a professional programmer, or any good at it really, i could have much difficulty
in pulling it off alone... and also because "au contraire" to all those enthusiasts on
"charitable" pirated donations of MSFT soft often do, tweaking it full... this trick can
reveal itself much more difficult.

I'll make you a deal RH, build a live CD with a similar compilar infrastructure like
mentioned, that could detect all my hardware on my command, so that i can install from online
all the source code packages i need and that could be tailored for this vector & parallel job
i want,... and install it by compiling in a more or less automatic fashion... well Gentoo,
Gobo are more or less good examples,... but i want more, i want vector&paralell FULL
optimization, and i want a very smart and capable installer.

For crying out loud, if isn't for something like this, what in the hell is supposed to be the
practical meaning of open source for non-developers !?.... what is the difference to MSFT!?...

Sorry i don't want to be rude, i know that you RH, must have all there is of source code in
Open format that exists on this planet in repository, and to be very frank i couldn't care
less about desktop vs server rhetoric... if it wasn't because of "popular" desktop we wouldn't
have neither Linux and RH... and i'm sure there is a market for desktop... lets see..

The only i care is about price; <$100:-to where i send the check; >$100 <$200 i'll see
Canonical and Novell first, i'm gonna think; >$200 i'll try to do it myself all alone first.

So do we have a deal ?

Ubuntu/Canonical not non-profit

Posted Apr 17, 2008 15:49 UTC (Thu) by rahulsundaram (subscriber, #21946) [Link]

Canonical is not a public company and afaik they haven't turned profitable yet. Besides they
have already asserted that their desktop is a means to get a foot in the door and sell it to
the enterprise eventually.

http://www.news.com/A-Linux-start-up-on-the-path-to-profi... 

Besides Red Hat continues to invest heavily in core desktop components. 

http://fedoraproject.org/wiki/RedHatContributions

I guess that wouldn't happen unless there is belief that there is profit to be made.

Ubuntu/Canonical not non-profit

Posted Apr 18, 2008 0:35 UTC (Fri) by szaka (subscriber, #12740) [Link]

70% of Ubuntu's business comes from server support: http://www.linux.com/feature/132575

And another thing...

Posted Apr 17, 2008 15:53 UTC (Thu) by bfields (subscriber, #19510) [Link]

"They're giving up on a big market segment by abandoning consumer desktops."

Or maybe they're approaching it in the way that makes sense for them, by first finding a few
niche markets that allow them to continue to support desktop work.

Not necessarily.

Posted Apr 18, 2008 3:48 UTC (Fri) by jd (guest, #26381) [Link]

Microsoft and, to a degree, Apple essentially own the metaphor of a desktop... but they don't
necessarily own the hearts and minds of home users. Just suppose there's a better metaphor, a
better way to present information to users, a better design that is not based around the
notion of a desktop with objects on it, but is based on some other paradigm entirely.

The desktop model has been useful in the past, but how many people actually use their computer
that way? I have no idea what alternative model you could use, but let's suppose there is one.
What then?

Well, it'd not competing with Microsoft on Microsoft's home turf, it'd not be playing catch-up
with Microsoft's capabilities, it'd not be attempting to replace a very firmly entrenched user
mindset. Rather, it would be seen as something altogether new (which is usually seen as good)
and incidently doing eveything the user wants (which is also good). I imagine that was part of
the idea behind XO, which has done a fairly decent job of coming across as something genuinely
different but is unfortunately still somewhat desktop in look and feel.

Could Linux try winning by entering a different race entirely? Fresco a.k.a Berlin seemed to
be looking at trying a different approach - fragmetable linked windows isn't exactly a desktop
concept. Xrooms also seemed to want to try going a different way. Both have long-since
vanished into history. Clearly, if that is how to win, it isn't going to be easy.

Not necessarily.

Posted Apr 19, 2008 3:16 UTC (Sat) by mmarq (guest, #2332) [Link]

"" Clearly, if that is how to win, it isn't going to be easy. ""

You forgot Fluendo with its Elisa, you forgot miroTV... LinuxMC... i can think of a couple of
business models around those applications only.

OLPC-OX is not a business model not by a half millimeter... its government subsidy... how can
anyone have mature and independent model based on subsidy ?

Yes there is too much fragmentation in Linux desktop yet. It would be nice to have cross the
desk the same look and fell. Common styles, icon sets, colors... whether KDE or Gnome or...
that is, there should be a basic theme model for everyone, yet a good and extensible engine so
that creativity would not be crushed.

But if big players don't bother, they have other politics, its up to the small fish to carry
the all load of progress into mainstream, though they tend to diverge a lot... and i say this
because servers are invisible to the population.

And people in Open Source are forgetting the most important of all, the source code must have
a meaning for non developers. No way in hell could a Linux "commercial" company sell me a
distro, not even for small server jobs. Lets not be hypocrites, its even true for MSFT, if it
wasn't for the more and more draconian ways of locking their 'soft' from unauthorized use.

I'm not even using one of the parallel to commercial distros (fedora,opensuse...) for that
matter... but every company "commercial" or not, could sell me intelligent ways of using
source code with great advantages...

Pitty people only think of MSFT models... until they get subsidized into oblivion.