Martin Schulze <[EMAIL PROTECTED]> writes: > > David Maslen wrote: > > The is a story on ./ at the moment with a writeup of a talk RMS gave > > in NZ. http://cantua.canterbury.ac.nz/~mpt26/stuff/rms/ > > > > One thing that seems relevant to this list is that RMS reportedly said > > using the MACH micro kernel was a mistake. Now does this mean it was a > > mistake because the micro kernel concept is bad? Or because it > > retarded developement of the hurd, and so linux took the free software > > innertia? > > As far as I know rms I believe that he's referring to the long time he > had to wait for the Mach micro kernel to appear. In the meantime Linux > has gone much farer and even the Debian project was started. Waiting > for the kernel to appear did hurt the Hurd project because people want > to do anything, they want to use anything.
I can think of two answers to this. One, Linux nowadays uses the same API as HURD (glibc2). The popularity of Linux means that a lot of code will port trivially to HURD. Every developer using Glibc is a developer not using Win32. Two, writing GNU-quality code takes considerable skill. Linux may not be GNU-quality, but the experience gained from Linux will help developers "move up" to the HURD. As M$ well knows, it's better to be second at doing something, because you learn from the first guys' mistakes (and successes). Even more so in the world of Free Software; you also learn and borrow from their code base. And besides, you just gotta go with the microkernel. ;-)
