The future of Linux
Andrew Stickney
I just discovered an interesting article analyzing the future of Linux:
The primary argument is that in order to identify the future of Linux,
one must not analyze it as an "OS" (since there are so many flavors)
but rather as a series of hardware/software environments and scenarios
(e.g. software development, mobile apps, server OS). The author seems
to cop-out a little in the end by saying the end-users are the
ultimate decision-makers as to where Linux goes in the future - it
could have had a stronger ending, IMO.
Let me know what you think,
Andrew Stickney
stick...@gmail.com
lunaz
friends) barely know what they are getting with their new computers.
They just want to do the basics. Because of this I don't think users
are the end all decision makers, whoever sells them their computer is.
I think people should realize there's more choices then use whatever
works best for them.
On Aug 11, 11:01 am, Andrew Stickney <stickne...@gmail.com> wrote:
> Hi MLUG,
>
> I just discovered an interesting article analyzing the future of Linux:
>
> The primary argument is that in order to identify the future of Linux,
> one must not analyze it as an "OS" (since there are so many flavors)
> but rather as a series of hardware/software environments and scenarios
> (e.g. software development, mobile apps, server OS). The author seems
> to cop-out a little in the end by saying the end-users are the
> ultimate decision-makers as to where Linux goes in the future - it
> could have had a stronger ending, IMO.
>
> Let me know what you think,
>
> Andrew Stickney
Michael Loftis
increases their profit margins in various ways and lowers their
support costs (atleast that's their perception). Same thing with HP,
Compaq. No one from any linux camp has offered them cash back for
buying their distro. Then for the big vendors there's the HUGE cost
of just installing any other OS on the scale they do things.
Cliff
late_rabbit
want to be branded an "offtopic" guy or similar.
I've taken some of the comments I've gotten to heart and feel like I
ought to find a different group or group with more interest in the
topics that interest me which are not strictly rendered IT questions
about particular software packages and particular hardware.
On the other hand, I've been reluctant to try since I don't want to
end up in a group of basically one person talking to himself.
If I were to go on, it would be in defense of the thesis, linux is as
linux does.
By that I mean, linux is most vital as a practice. We should question
the value of replacing windows on the desktop in the sense that linux
some day becomes sufficiently commodified that its widely adopted.
Whether the primary value of linux is whether it can be commodified is
the real question. A key part of the power and appeal of linux is
that its existence proves that a computer science student can spend a
long winter in Helsinki and make something so wonderful.
Anyhoo . . . .
robert bowman
> Whether the primary value of linux is whether it can be commodified is
> the real question.
Whe Linux is commodified, it will be time for détournement.
late_rabbit
word. Nice.
Michael Loftis
big deal, but not that big of one. It's the application ecosystem
thatlives on top of that kernel (and most of themalso live on top of
MANY other kernels) that makes it useful and attractive to persons.
RedHat, by and large, is commoditized. Ubuntu and Debian are a ways
behind RH. The Linux Kernel itself, well, it's useless without all
the other stuff around it.
Bill Day
this one.
First, although the original article uses the rubric Linux, it is really
about GNU and Free Software. Yes, Linux is in one sense "just a
kernel," but it is a kernel that came into being at just the historical
moment when a Free kernel was necessary to complete a Free operating
system. The GPL, more than any particular piece of software, was the
revolutionary development. However, in part by historical accident,
Linux has been far more than "just a kernel."
As for commoditization, Linux systems are niche products, but it is my
impression that the existing user base, as opposed to the mass of
indifferent Windows users, secretly likes it that way. Caviar and foie
gras may not appeal to the same palate as Captain Crunch, nor the
Bentley sell to the same crowd as the Buick, but mostly the former do
not seem overly disturbed by the unlikely prospect of winning over the
latter. A challenging, complex, open, flexible, powerful OS does rely
upon a certain snob appeal. Fundamentally, trying to turn an objet
d'art, or better, a specialized tool, into a commodity may be neither
possible nor desirable. Maybe you can't turn a Maserati into a Mazda,
but do you really want to?
My impression is that Linux systems are unusual in that a higher
proportion of their users are developers than most other operating
systems, a fact that militates in favor of software that is at once more
stable and more flexible than that of closed source competitors. So
long at there is a sufficiently large community to drive development,
how important is it to dominate the marketplace, at least from the
perspective of those of us who do not make our livings either selling or
supporting software?
Naturally, I think more widespread adoption of Linux would be a good
thing. But it is not the only thing, and probably not even the main
thing. The main thing is having and developing a robust system that
gives everyone full access to the system and the code. Clearly, such a
system is only for some users, but some users are very much for such a
system. To answer Cliff's point about the "blame game," if you are your
own builder and systems administrator, you have no one to blame but
yourself when you screw things up. And thank God for the community when
that happens.
late_rabbit
Michael Loftis
open bsd, and some others that escape me. The bsd are more open source
than linux some would argue since they're not encumbered by the GPL
and its all changes must belong to the public stance.
It wasn't just the fact it was free as in beer nor open as in
available. There were many factors, most of which I don't even know.
Sent from my mobile device
robert bowman
> It wasn't just the fact it was free as in beer nor open as in
> available. There were many factors, most of which I don't even know.
I think the BDSs sufferered from politics. The Jolitz' 386BSD forked into
FreeBSD and NetBSD over internal dissatisfactions. When de Raadt, a younger
version of RMS when it comes to personality, started OpenBSD when the NetBSD
community ejected him.
The dilution of energy and talent didn't help, plus the general confusion of
having three different, more or less the same, branches. Linux was more
fortunate in that the distros branched and achieved their own identities but
they all are known collectively as 'Linux'. There is brand recognition. RMS
can state frequently, loudly, and correctly that it is 'GNU/Linux' but he
lost that war a long time ago.
I don't know if Torvalds is mellower than de Raadt, Tanenbaum, and Stallman, a
more skillful manager, or just benefitted from another simple twist of fate.
Bill Day
the GPL. Many people who are quite willing to donate to the common
good for free in the confidence that others will reciprocate may be
less eager to donate their code for free if it is going to inure
largely to the benefit of the bottom line of, say, Apple Computer,
with no reciprocity. (I am aware of the irony of my posting from my
iPhone.)
Sent from my iPhone
robert bowman
> I think it may be possible to underestimate the motivating force of
> the GPL. Many people who are quite willing to donate to the common
> good for free in the confidence that others will reciprocate may be
> less eager to donate their code for free if it is going to inure
> largely to the benefit of the bottom line of, say, Apple Computer,
No offense to the present company, but I'm not a lawyer and I never had any
patience for RMS's legalisms. If I donate code, it's just like giving a guy
on the street a few bucks. It's gone, I'm done with the transaction, and
whatever ripples it may cause aren't my concern.
Bill Day
surprise perhaps that, being a lawyer rather than a developer, I think
more like the one than the other. However, even in the case of the
guy on the street many people may be selective, e.g. I am more likely
to give to a woman with kids than to a man. Or is there a better way
to explain the otherwise surprising interest of people in contributing
to the more rapid development of Linux than the BSD's?
Sent from my iPhone
late_rabbit
lawyers' work. :)
robert bowman
> Or is there a better way
> to explain the otherwise surprising interest of people in contributing
> to the more rapid development of Linux than the BSD's?
If yuo're talking about the kernel, I think it's a brand recognition and
momentum thing. The BSD's tend to lag in drivers for wireless cards, etc.
Less interest, fewer people who are willing to do the work == less interest,
fewer people.
On the other hand, look at the Perl and Python communities. Larry Wall's
personal feeling is 'here it is. have fun', with his tongue in cheek 'Poetic
License' and the 'if you really want to GPL this, knock yourself out' clause.
Python had even a more ambiguous license situation. The MIT or X License is
also very permissive.
On the other hand, the trolls had a licensing scheme for Qt that required a
corporate lawyer to decipher, which to some degree limited its use.
It might be the circles I've travelled in, but most of the coders I've worked
with were big on code reuse without reading the fine print, and kicked bug
fixes back into the gene pool. For instance, we used to use Expect. When I
found a particulary nasty sigsetjmp bug, I fixed it and reported it to Don
Libes so it would go into the next Expect release. If the license had
precluded commercial use, that wouldn't have happened.
robert bowman
> Lawyers here are more interested in coders' work than coders are in
> lawyers' work.
In a way, it's ironic. Both professions depend heavily on precise parsing of
complex expressions written in arcane language, but I don't think there is
much crossover. I wouldn't say coders are legally lax, but not too many find
EULA's fascinating as long as the damn thing will install, and should a peice
of code fall off the turnip truck, well that's the code reuse that everyone
says is a good thing.
late_rabbit
lawyers, but that's how your comment sounds. Both hack systems. Well
--- maybe the lawyers who get excited about M$ EULA's aren't much for
"hacking" as such. Eye of the beholder. Lawyers hack systems in the
"real world" or are caught in the matrix. I haven't decided
which. :)
Michael Loftis
On Mon, Aug 31, 2009 at 8:22 AM, robert bowman<bow...@montana.com> wrote:
<snip>
> On the other hand, look at the Perl and Python communities. Larry Wall's
> personal feeling is 'here it is. have fun', with his tongue in cheek 'Poetic
> License' and the 'if you really want to GPL this, knock yourself out' clause.
> Python had even a more ambiguous license situation. The MIT or X License is
> also very permissive.
This is why if you look you'll see python in the guts of more than a
few commercial games for example.
>
> On the other hand, the trolls had a licensing scheme for Qt that required a
> corporate lawyer to decipher, which to some degree limited its use.
Just so everyone knows when he says trolls he's (I hope ;) ) meaning
Trolltech, now Qt Software. But that was/is very true the Qt license
really limited what it could be used for, partly because no one could
make heads or tails of it. The other issue was there were (and are)
many other competing libraries out there in that space.
> It might be the circles I've travelled in, but most of the coders I've worked
> with were big on code reuse without reading the fine print, and kicked bug
> fixes back into the gene pool. For instance, we used to use Expect. When I
> found a particulary nasty sigsetjmp bug, I fixed it and reported it to Don
> Libes so it would go into the next Expect release. If the license had
> precluded commercial use, that wouldn't have happened.
Yeah and that's the general consensus of Unix/Linux/BSD/etc coders.
The license might allow you to take it in and keep it, but no one
REALLY wants to maintain 100x bug fixes on top of the public domain
code, so bug fixes, and even sometimes wholly new features, get
committed back into the trunk even if the code is being commercially
used. Windows coders tend to have a different view because they're in
an ecosystem that doesn't have a lot of free software - though that's
starting to change largely because of portability shims like Cygwin
and wx* that run on many platforms. It's really a product of your
raising as it were. In an ecosystem that sharing isn't the norm, they
don't learn to share. In our ecosystem where sharing and community
driven effort and development is the norm, we naturally tend to share
more.
Michael Loftis
>
> Haha ... I would not have thought of coders as more practical than
> lawyers, but that's how your comment sounds. Both hack systems. Well
> --- maybe the lawyers who get excited about M$ EULA's aren't much for
> "hacking" as such. Eye of the beholder. Lawyers hack systems in the
> "real world" or are caught in the matrix. I haven't decided
> which. :)
Legal systems are all hundreds of years old without much chance of
being replaced...They're the huge monolithic Cobol or Fortran app
equivalent. It's not the lawyers fault the 'code' they write sucks,
they're being forced to deal with bad parsers and an antiquated
system! :)
Michael Loftis
we in the coding world often reduce to RTFM.
Tho... telling a judge or another lawyer to RTFM in a court of law
probably wold NOT fly well, once you explained what you meant...I'd
love to see the results though if it were tested LOL
robert bowman
> Windows coders tend to have a different view because they're in
> an ecosystem that doesn't have a lot of free software - though that's
> starting to change largely because of portability shims like Cygwin
> and wx* that run on many platforms. It's really a product of your
> raising as it were.
There was a bifurcation somewhere along the way. There was quite a bit of free
software for CP/M, Commodore's C64 and VIC-20, and the various Arari's. Even
DOS had freeware. I did a little work with DJ DeLorie's djgpp project, mostly
bug fixes as I tried to port Midnight Commander to DOS. That was mostly a
learning experience for me, since MC was a clone of the original DOS Norton
Commander. Mumit Khan had inherited the MinGW project and there was some
overlap. Anders Norlander headers allowed writing Win32 GUI apps in C. I'd
also done a little bit with that project, chiefly a clean room implementation
of DirectX headers. Miguel de Icaza also passed through at random moments,
although he was already heading in a different direction. There was a lot of
overlap, but Cygnus was a little later than djgpp.
The energy was there, but not much made it to Windows applications
programming. The last I knew DeLorie had a day job with RedHat, and de Icaza,
of course, is driving mono. I've no idea where the rest of the people have
wandered off to in the last 15 years or so.
I think part of the problem, which has also hampered Linux to some extent, is
the interests of the people who get involved in open source tends towards
tools and systems programming rather than applications. Compilers and
interpreters are cool, as are WM's and so forth. Stuff hoi polloi use, not so
much so. For instance, Quicken was the killer app for Windows at one point,
and I'd look at developing something like that as the fate worse than death.
Of course, if a suitable paycheck is dangled in front of you, you can work up
an interest, but that takes care of the freeware deal.