>Well, for what it's worth, _I_ appreciate being able to build a kernel
>sufficiently generic that it'll work on any SPARC, or at least any
>SPARC that NetBSD works on at all...

Count me as one who appreciates the single-kernel-for-all-sparcs model as
well; I'm contemplating an upgrade now from an ancient Sun4/330 to an LX
(sun4 to sun4m) under SunOS 4.1.4, and it's not just a hardware swap &
reboot (as it should be!). One good reason to go NetBSD is that it looks
mature enough, and I can play hardware swap games with a production system
like that.

I have no doubt that if we ever get around to cutting a bootable CD-ROM
distribution, the single-kernel-for-all-sparcs will also make that lots
easier. The payoff is worth the effort.

	Erik Fair <fair@clock.org>

P.S.	Please do not construe this as meaning that I don't like the
ability to remove all the code that my system will not run by reconfiguring
a new kernel; that optimization is quite useful, too.