[lkml]   [1999]   [Feb]   [13]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
SubjectRe: Kernel interface changes (was Re: cdrecord problems on
In article <>,
Nat Lanza <> wrote:
>I'd rather pay $20 to
>have my fancy new soundcard work than wait months for someone to write
>a free driver.

Been there, done that. I would saved US$7 by buying a cheaper sound
card but I would have had to pay US$30 for the OSS driver. In the end
I saved US$23 by buying the more expensive but supported card. Now in
2.1 the cheaper card is supported for free too.

I used a number of commercial X servers in 1996-1997 to run some video
cards unsupported by XFree86 (I think they were Matrox Millenniums or
something). The business case for the cards was that they were about
ten times faster than anything at their price point on the market (which
was true). This was something of a disaster from day one as the cards
never really worked. We started with one card for demonstrations to
clients but quickly replaced it with a Mach64 card because a slow demo
was better than crashing in the middle of a demo. In 1997 I replaced
_all_ of the commercial servers with XFree86, when XFree86 had solved
the problem of making a Matrox card work. We wasted hundreds of dollars
on that commercial X server and we had to hide it in the closet when
clients came to the office lest they see just how crappy it was and think
that it was the fault of _our_ software!

>I like OSS just fine, but I don't run Linux on my
>machines to get into a state of free software religious purity. I run
>it so I can get some work done.

Unfortunately this is what happens if you let non-free software get into
your computing infrastructure and if you don't keep your infrastructure
up to date. Alas, almost everyone has done this and now we are beginning
to understand the true costs for the first time.

>Making life difficult for binary modules is great if what you really
>want is to have a pure, all-open-source hobbyist environment.

Binary modules are fragile to begin with. They'll easily break even if it
was a stated goal not to break them. Life is, was, and always will be
difficult for people who produce and distribute only binaries. If you
want stable binaries go talk to Sun and see if they can nail down the
Java spec for you. ;-)

>Unfortunately, not everyone who uses Linux these days is a hobbyist or
>even wants to be. Some of us chose it because it's a better system
>that we can use for our work. If I can contribute something to the
>community through porting our NASD research code to Linux, that'd be
>great. But it's not my primary goal. If that makes me a bad citizen in
>the Linux world, so be it. As the MIT folks have said repeatedly,
>there are several other high-quality free Unixes out there.

Lately I've noticed a downward trend in the overall quality of Linux, both
kernel and some distributions. Red Hat's release 5.[01], for example,
was nothing less than a disaster, although 5.2 seems to have recovered
(at least to the level of some of the better, earlier Red Hat releases).
Linux 2.2 came out the door about a month too early IMHO--there are
some _serious_ bugs still floating around in there that should have been
squashed, and the '2.2.0-preXX' series was _way_ too short and had too
many new features coming in at the last minute. It takes a week or two
just to run a minimal integration test suite these days, and those preXX
kernels were coming out a few days apart.

Zygo Blaxell, Linux Engineer, Corel Corporation, (work), (play). It's my opinion, I tell you! Mine! All MINE!
Size of 'diff -Nurw [...] winehq corel' as of Sat Feb 13 15:14:00 EST 1999
Lines/files: In 10558 / 114, Out 6460 / 77, Both 14451 / 162

To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to
Please read the FAQ at

 \ /
  Last update: 2005-03-22 13:50    [W:0.053 / U:1.896 seconds]
©2003-2018 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site