On Tue, 27 Oct 1998, Dennis wrote:

> If we do sell a "low-end, source" product, we'll do it for Linux, because
> the market
> is 50X greater. Why deal with a piddly market like NetBSD at all?

That, in fact, would probably satisfy the needs of a lot of the
folks here, and we'd encourage you to go ahead with that. If you
have a source product for Linux (adding good documentation would
help a lot), you don't have to worry about the NetBSD market,
because someone's likely to port the driver, and maintain it as
you update the Linux driver. (We already do this with several other
drivers.)

In the end, I think a lot of the disagreement here is due to
differing models of software. From the NetBSD point of view, you
have several products bundled together: a serial card, a driver,
frame relay support, bandwidth management support, and so on. You
add frame relay support to an OS by putting it in to your driver
and adding the driver to the OS, whereas the NetBSD philosophy
would be to put generic frame relay support into the OS, and have
it work over any serial driver. That happens to conflict with your
model of how to make money from your efforts, and there's no easy
way to resolve it. That's unfortunate, but that's life. You'll have
to go your way, and we'll have to go ours.

As far as drivers for NetBSD, I'd recommend porting to 1.3.2 right
now, which is almost certain to give you compatability with the
entire 1.3.x line for as long as it exists (including releases not
yet done). Then, when current turns into 1.4 alpha, port your driver
to that, which should give you compatability with the entire 1.4.x
line for all time. It's hard to recommend porting to current because
current tends to change a lot, and working on one version of current
doesn't mean you'll work on another. 

cjs
--
Curt Sampson  <cjs@portal.ca>  604-257-9400    De gustibus, aut bene aut nihil.
Any opinions expressed are mine and mine alone.
The most widely ported operating system in the world: http://www.netbsd.org