Possibly offtopic : Binary only driver

[Paul Iadonisi]

On Sun, 2004-11-21 at 23:04 +0000, Mike Hearn wrote:
> On Sun, 21 Nov 2004 17:46:01 -0500, Sean wrote:
> > You haven't made a good
> > case that the kernel developers should change anything about the way they
> > handle these issues.
> 
> The case I'm making revolves around Linux being an OS people can use to
> get things done, rather than a work of art or a demo of open source
> principles. If you think the kernel is better off as a flagship of purity
> rather than a production desktop kernel then fine, I have nothing to say
> about that.

  I think you are neglecting another possible alternate future for Linux
that I believe is actually much more likely.
  In a similar discussion not long ago on either this list or fedora-
test-list, I reminded someone of a similar mind as you the history with
SCSI card vendors.  It used to be like pulling teeth to get SCSI card
vendors to release specs for their cards, much less GPLed drivers.  But,
then something happened that they didn't expect[1]: Linux started taking
off (and some cases, taking *over*) in the server room.  It wasn't long
before SCSI vendors were falling over each other to get their GPLed
drivers rolled into the mainline kernel.  Xircom (now owned by Intel)
comes to mind as a prime example.
  Fast forward to a few years in the future and I believe your earlier
false dichotomy here --

On Sun, 2004-11-21 at 21:10 +0000, Mike Hearn wrote:
> That's why I said "primarily useful". Obviously open source code is
> good
> for society in general, that's why we're all here! My point was that
> it's not so directly useful to non-programmers that they'll say "Oh
> well
> I'll not play 3D games anymore" en-masse. Not going to happen

  fails to identify a common catch-22 that can, and has in the past
(like in the SCSI case) been broken.  There is no market for games on
Linux so there is no market for Linux among gamers.  As Linux desktop
systems take off more and more (and that, I believe is inevitable) in
business and non-gamer home use (and I know many non-gamer home users,
BTW) you will start to see more demand for games on Linux.
  *Someone* will produce a decent graphics card on Linux and provide
Free Software drivers that will end up in the stock distributions of the
Linux kernel and X.org.  If you think that's unrealistic, search the
linux-kernel archives over the past month or so for a discussion
initiated by hardware vendor engineer who was interested, and had buy-in
from his management, to produce a completely open design with Linux in
mind specifically.  It was (is?) going to be 2D initially, but the
possibility of 3D was not out of the question for the future.
  So eventually there may be a non-nVidia, non-ATI graphics card with
3D-acceleration on Linux supported through Free Software drivers.  This
will make it possible for freedom-conscious distributions to bundle
these drivers making this vendor's card the only well supported 3D-
accelerated card on Linux out-of-box.
  Any company migrating most of their desktops to Linux will suddenly
see an opportunity.  Some additional percentage (1-2% maybe?) of their
desktops that have been held back to that *other* operating system
because their sales engineers need high performance 3D graphics to do
their presentations will now be able to switch to RHEL5 (or WBEL5.0, for
the really cost conscious) because it 'just works' out of the box with
this new 3D-accelerated card.  These 1-2% need hardware upgrades,
anyhow.  Now that this company can save money on software licensing, due
to the ability to migrate to Linux, these upgrade have become
affordable.  (Presuming, of course, that there are no Windows apps
actually holding back these specific upgrades, just lack of out-of-box
3D-accelerated support.)
  And then who's to say that Sony Playstation 3 game vendors won't begin
to see an opportunity as well?  "Gee, we're already developing on
Linux," (as the PS/3 supposedly will be based on), "Linux is now gaining
popularity, and also has an out-of-box working 3D-accelerated card.  So
we have the Linux expertise, let's get our games working on desktop
Linux!"  (Yes, I know about restrictive console vendor contracts, but
that could all change if the economics do.)

On Sun, 2004-11-21 at 15:40 +0000, Mike Hearn wrote:
> Well this is just a generalisation of "all software should be open
> source". I tend to agree with that, it would be great if that were
> true.
> 
> But it's not true, and there are no signs of it suddenly becoming true
> anytime soon. I'd also rather people open sourced their code due to
> the
> social benefits (everybody being able to fix bugs, share knowledge,
> implement new system-wide optimisations etc) rather than because the
> kernel made it a total pain in the ass to do otherwise. That's
> coercion
> not persuasion.

  And when this 'third' graphics card vendor I mention above *owns* the
market for 3D-accelerated cards on Linux, nVidia and ATI will wake up
and start to consider the benefits of releasing their code.  They may
not do it right away, but they will be forced consider the *economic*
reality that will become evident.  It's not the case now, but,
paraphrasing Jon 'maddog' Hall, "it's inevitable".
  Nobody is coercing anyone, as you imply above.  If Linux becomes
wildly successful on the desktop (with or without nVidia or ATI), which
I believe will happen, and you want to play, you play by the rules.
That's no different than any other operating system.


On Sun, November 21, 2004 1:43 pm, Mike Hearn said:
> I think you need to talk to the nVidia engineers and/or Alan Cox, who
> has
> said in the past (I think) that he can't find a way for them to open
> their
> driver sources without suffering serious consequences. It's not a
> simple
> matter of patents and legal problems. It's a matter of economics.
> 
> Anyway, this whole point is silly: nobody should be *forced* against
> their
> wishes to open source their code if they don't want to. If open source
> development really is better than the old way, then rational people
> will
> become a part of it over time if they can.

  Agreed.  But nobody is *forcing* anyone.  The binary-only vendors have
made their own beds ... let them sleep in it.  And there's no reason
whatsoever to make it comfortable for them to not join.  Personally, I
don't think ndiswrapper is good idea for that very reason.  It gives
network card vendors a reason to be lazy and not port their drivers and
provide them to the Linux kernel developers for inclusion.
  When nVidia and ATI come and join us, I have no doubt it will be for
economic reasons.  That's fine.  Provided they come with an
understanding of the terms of participation ... and this is to
contribute code under the appropriate licenses, not to try and change
the Linux way of doing things.  Code is the only currency of the
meritocracy that is Free Software.  And you do have to 'buy' your way
in.


1. This is key.  Linux keeps on achieving things that ISVs, IHVs, and
   pundits don't expect.  It's an ingrained characteristic of Free Software
   in general.  When the pundits say, <random-free-software-project> won't
   achieve <some-random-achievement> any time soon, it's obvious they are
   saying that they don't believe it will *ever* achieve it.  The Free Software
   projects keep proving them wrong, but they never learn.  The lesson: *expect*
   to be surprised.

-- 
-Paul Iadonisi
 Senior System Administrator
 Red Hat Certified Engineer / Local Linux Lobbyist
 Ever see a penguin fly?  --  Try Linux.
 GPL all the way: Sell services, don't lease secrets