GNU License for Hardware

[mark at pc-intouch.com]

On Sat, 16 Oct 1999, Tom Hull wrote:

> Finally, why should we trivialize the kernel of any OS as an "only
> thing"? If kernels were so easy, one would think that GNU would
> have long ago released one. But in my experience kernels are not
> so easy, which is precisely why I think that anyone who seriously
> intends to build an Operating System has to tackle the kernel
> first. Perhaps some respect is also due to the people who have

I disagree with the statement about tackling the kernel first, especially
in this context.

FSF built the programming tools, applications, and libraries first, and
then the kernel (which isn't finished yet, and might never need to be
finished considering that there are plenty of Free kernels out there that
(a) interoperate with GNU stuff and (b) actually work).

The result of this is that today it's not that difficult to port GNU tools
and apps to run on any Unix-style kernel, which means that we could, if we
were so inclined, create GNU/Solaris or GNU/HP-UX or whatever.  In theory,
the proprietary Unix vendors could realize that it's not profitable for
them to try to maintain proprietary tools for their platforms when they
could just port the GNU stuff, which usually works better anyway. (Doesn't
SCO ship a CD of GNU software with one of their distributions?  I remember
seeing that somewhere.)

If FSF had started with a kernel, they might very well have finished the
kernel first and then designed all their tools to work specifically with
their kernel.  They'd lose the sort of gradual infiltration of GNU
software into proprietary Unix packages that we're seeing, and end up
writing an OS that works very well but doesn't take advantage of the
potential support of the proprietary vendors.