On Tue, Oct 27, 1998 at 12:48:02PM -0500, Dennis wrote:
> At 11:49 AM 10/27/98 -0500, you wrote:
> To paraphrase your own words back:
> 
> >No, but you are demonstrating your lack of understanding, as are most of
> >the other people in your vendor camp.  Any OS that tries to make
> >everyone use it regardless of their requirements doesn't have a
> >marketing plan worth implementing.  If you can do without a complete;
> >extremely portable; highly interoperable; well designed, stable, and
> >fast; freely available and redistributable; highly standards compliant;
> >UNIX-like operating system then you aren't a candidate to use our
> >product. 
> 
> But they can, because they can use FreeBSD, which fits your
> description, at least in the i386 world which is all that matters to us.
> 
> >> Most of our customer base are BSD/OS and NetBSD convertees. 
> >
> >That's odd.  One of the products I know that incorporates support for
> >your products is based on BSD/OS and you seem to continue to advertise
> >BSD/OS compatability on your web page....  Why yould you support BSD/OS
> >and then try to convince BSD/OS users to use FreeBSD?
> 
> Because we are not supporting BSD/OS 4.0 and beyond.. Most of the convertees
> converted because they saw no reason to pay for BSDI when they could get 
> the same functionality with FreeBSD. Not our doing, but a natural migration.
> 
> Clearly your words indicate that you are not interested in having a widely 
> popular OS, which certainly is a major criteria for a vendor.

Now, that's kind of obnoxious.  I mean, "clearly your words indicate that you
are not interested in having widely popluar hardware, which certainly is a
major criterion for an operating system project trying to come to an
understanding of some type with you, a hardware vendor."

I am glad that you're considering porting your software to NetBSD.  I
encourage you to do so, and I expect that if you do do so, people will buy
your cards and use them with NetBSD.  But it's just plain silly to get in an
ideological argument with people who have a different model for software
development and support than you do, and then go off in a huff because _they_
wouldn't do things the same way _you_ would.

To begin with, vocal users on the mailing list don't speak for the project;
if some kind of architectural changes to better support binary-only device
drivers would be helpful to you, for example, you might want to kick the idea
around with the project's management, core@netbsd.org.  I don't think anyone
would go out of his way to make your life hard, that's for sure.

We run on far more hardware than FreeBSD or Linux does, or probably ever will.
A substantial fraction of those systems are at the "high end", like the big
AlphaServers.  Another advantage of porting to NetBSD is that you can sell the
same PCI device driver to customers with a PCI x86 box, Alpha, SPARC
(presently -- hi, Eduardo! :-), PowerPC machine (Macintosh or elsewise),
without having to maintain it N times.  If you were inclined to split apart
the hardware and link layer parts of your driver, we'd probably maintain the
hardware part _for you_, in our source tree.  To a lot of the people you're
talking to, that seems like such a big win for you that they're having
trouble understanding why you don't see it that way; that's all.  Nobody (I
hope) is trying to push you around about how to sell or not sell your hardware
or software.

To your original question: I suggest you port to -current as of the latest x86
snapshot (I can extract a source tree as of that date and send it to you if
you like) and release binaries now; then you whould have minor porting work to
do to be able to recompile for x86 1.4.


-- 
Thor Lancelot Simon	                                      tls@rek.tjls.com
	"And where do all these highways go, now that we are free?"