Home Messages Index
[Date Prev][Date Next][Thread Prev][Thread Next]
Author IndexDate IndexThread Index

Re: The question about adding graphic facilities in future versions of Linux Kernel

  • Subject: Re: The question about adding graphic facilities in future versions of Linux Kernel
  • From: Roy Schestowitz <newsgroups@xxxxxxxxxxxxxxx>
  • Date: Wed, 07 Jun 2006 17:12:33 +0100
  • Newsgroups: comp.os.linux.advocacy
  • Organization: schestowitz.com / MCC / Manchester University
  • References: <1149651674.505537.107270@h76g2000cwa.googlegroups.com> <2OednUc-odEkaRvZnZ2dnUVZ_sadnZ2d@comcast.com>
  • Reply-to: newsgroups@xxxxxxxxxxxxxxx
  • User-agent: KNode/0.7.2
__/ [ Michael B. Trausch ] on Wednesday 07 June 2006 16:46 \__

> Justin wrote in <1149651674.505537.107270@xxxxxxxxxxxxxxxxxxxxxxxxxxxx> on
> Tue, June 6 2006 23:41:
>>
>> In order to achieve the goal that we all of us have in mind ,ie Linux
>> becoming a mainstream O.S - used by a majority of users , i think it
>> should have a very powerful edge over its competitor.
> 
> That is an interesting way to state the "goal" that you assume all of us
> have in mind.  Linux _is_ mainstream, today:  You can say the word Linux
> and most people will know that you're referring to something that they
> haven't used yet.  :)  Mainstream does not necessarily mean that it has the
> majority or plurality of the mindshare out there.  The fact that Linux is
> getting more coverage in magazines geared to the teenagers that provide
> technical support for their families means that that there are more people
> that are using it -- and liking it -- then you probably think there are.
> 
> Also, you seem to be missing that Linux already has a very powerful edge
> over its "competitors" -- it is completely open.  Since the source code is
> available, anybody that has the desire to learn its programming interfaces,
> can.  Not to mention that many of the applications that are used on a Linux
> system are also free, open-source software.  This is a significant "edge"
> over competition.  However, users in general do not yet understand the
> benefits of open source software.


Part of this should be attributed to the flawed contention that popularity
implies divinity. To those who have always dealt with exectable files, the
'other' option seems rather non-existent, let alone the possibility of
tinkering (or 'hacking'). With complacency comes acceptance. With diversity
comes better understanding, choice, and freedom (or Freedom if you like).


> If Linux needs anything else in the way of an edge, it needs people who are
> willing to be educators for the system, people who are willing to
> distribute it, people who are willing to show people what it really is.
> The problem that stands in the way of wide-spread adoption (at least, in
> the US) is two-fold: People have crazy misconceptions about the system, and
> interoperability is something very close to -- but not quite -- 100%.
> Also, DRM technologies need to be developed for Linux that allow people to
> buy online music, because that is the way the law is structured here with
> regard to music.  It is very easy for users to buy CDs and back them up by
> ripping them to music collections, however, it is not so easy for users to,
> say, use iMesh and download tracks that they want to hear without buying
> the whole CD.


It is rather intersting to me that you too were drawn to the DRM conundrum. I
seem to associate it with just about any joining of hardware and some
mysterious binaries.


>> But there seems to be major differences in the industry,  which left me
>> really disilussioned and which i thought i will share with you all.
>> Linux and and a lot of other programmer from around the world would
>> like to incorporate device drivers from the Nvidia and  Ati Radeon
>> graphic card companies(Only then linux will be able to give equal or
>> better visual experience compared to the AERO user interface of VISTA)
>> in the coming version of Kernel.
> 
> There was just a rather large thread on the kernel mailing list about
> putting graphics primitives and the like into the kernel.  I, for one, do
> not agree with such a proposal.  The kernel should be an interface to the
> hardware.  The 3D performance on nVidia is quite nice -- I can speak to
> that because I had a computer with such a card in it -- and the 3D
> performance for the Intel video in my laptop is quite nice, as well.  There
> is no need for graphics primitives and drivers in the kernel itself.  What
> there *is* a need for, though, is more well-written drivers that can take
> advantage of other cards that have 3D hardware acceleration capabilities.


3D in the kernel?!?! Never. Any sane person would have shunned this proposal
before it was even uttered in its entirety. As Tanenbaum recently stressed,
the kernel needs to be compartmentalised rather than monolithic. If you
shove anything into the kernel level, you make it more susceptible to
breakage than it already is. Testing is slowed, security is jeopardised, and
bloat is foreseen. Why, for instance, should a Web server with the Linux
kernel need to spend resources or volume on 3-D support?


> X11 is quite a stable platform, and the interfaces that run on top of it
> are
> very usable today.  The only lagging issue is 3D support for a lot of
> chipsets -- and, funny that, it is also a problem with low-end hardware in
> Windows systems.
> 
>>
>> But since these device drivers are NOT open source ,  and its very
>> unlike that they go open source in the near future , GNU foundation and
>> a lot of other programmers have opposed this move. This is a very
>> saddening fact , because as I understood this was a clear case of
>> idealogical clash. The GNU group wants kernel to remain completely open
>> source with not even a single bit of closed source code(If i may say
>> so)
> 
> You are right:  The nVidia and ATI drivers that have the "good" support for
> 3D acceleration are not open source.  However, the system that they run on
> top of, is.  At least for me, that is quite fine.  Would I use an open
> source driver if it were available?  In the case of nVidia graphics cards,
> the answer is "no," since there is one available, but the graphics
> performance on the proprietary driver is a lot better.  So what if the
> hardware driver is closed-source?  That might make it a bit harder to
> debug, sure.  However, it works (mostly) and that is fine.


If Linux was ever to be embraced as the platform of choice (the one and only)
for gaming and environmental design, something's gotta give. The industry
would have to lean over or new brands emerge, as I opined in the previous
message in this thread. I hope it makes a little sense...


> It is kind of like debating religion:  You have people that are
> fundamentalists on both sides of the debate, and then you have the people
> that are somewhere in the middle, considering arguments on a case-by-case
> basis, and coming to conclusions that way.  Personally, I think that people
> should be looking at using open-source on a case-by-case basis.  Now, if
> there comes a time where, say, an open source nVidia driver comes into play
> that has support for the hardware that rivals what the closed-source driver
> does, then it would be worthy of consideration.  However, that is not the
> case at this time.
> 
> Of course people are going to oppose closed-source software solutions for
> things.  However, I think that if the driver is there, and you are allowed
> to use it (which, in the case of nVidia, you are), then what is the
> problem?  Sure, you cannot tinker with it, and sure, I would be in favor of
> nVidia opening the source code for their driver, but if they do not want to
> do that, it *is* their prerogative.  After all, *they* wrote that code.
> Don't like it?  Then write one of your own!
> 
> Also, there is one additional flaw to your argument:  The GNU people do not
> control the Linux kernel.  The GNU people have their own kernel that is
> part of the GNU system, called the Hurd.  It is, last I checked, not
> functional.  Even if it were to become functional, I do not see it becoming
> a major player:  Everybody that is running GNU userland software, for the
> most part, is using the Linux kernel (with some exceptions).  The GNU group
> probably would not be tolerant of closed-source binary drivers for their
> system, and that is *their* prerogative.  However, if it were not for
> fanatics such as the FSF, we wouldn't have the happy medium that Linux
> provides for, today.  If you were to ask me, I would tell you that the
> happy medium is what is required to make the world go 'round, but that is
> just my own personal 2¢.


The FSF are, in my opinion, no fanatics and they rarely exaggerate either. If
you allow vendors to throw some binaries at your direction, you are asking
for more of the same. If you decline, you give them a run for their money
(or source code). It's bad enough that some companies (even Google) are
already throwing proprietary binaries at your direction, often gaining
popularity at the expense of openness. That, of course, is just my personal
opinion.


>> So what is your opinion , When VISTA finally arrives dont you think in
>> order to give a tought fight for its stuff dont we need to do the best
>> we can ensure that Linux becomes the global leader in O.S . Share your
>> thought here
>>
> 
> Vista will come, and it will do whatever it is going to do for those people
> who make the choice to adopt it.  It would _never_ be a good thing for only
> one system to be the system that everybody is running:  That would stifle
> any type of innovation.  However, I do see a time coming where Microsoft is
> going to be running a plurality or so, but not necessarily the majority of
> machines.  It may even be in the next five years, but I would say that it
> would be more likely to be in the next fifteen.  Microsoft cannot stay on
> top forever, and lately, they have been making massive blunders that have
> made even their most faithful people start looking at alternatives.


This seems evident. I don't want to turn this into a Microsoft-bashing
paragraph (admittedly, I have this tendency).


> We do not need to ensure that Linux becomes a "global leader."  The
> software
> is well on its way to going there.  What is missing?  The people who have
> the drive to help those see that the software is a realistic choice for
> them already -- and will be for years to come.


Some of the best advocacy campains are demonstration of practical success,
using live personal case studies. Just as they say "the best revenge is
living well", the best way to persuade people to upgrade to Linux is to show
them the positive impact on Linux on people's lives.

Best wishes,

Roy

-- 
Roy S. Schestowitz      |    "Disk quota exceeded; sig discontinued"
http://Schestowitz.com  |  GNU is Not UNIX  ¦     PGP-Key: 0x74572E8E
  4:50pm  up 40 days 22:23,  11 users,  load average: 0.65, 0.87, 0.92
      http://iuron.com - proposing a non-profit search engine

[Date Prev][Date Next][Thread Prev][Thread Next]
Author IndexDate IndexThread Index