It was thus said that the Great Liam Proven once stated:
I personally am not so concerned with the little nooks and crannies
that have been left behind -- except that I like to own and play with
some of the vintage computers that I can now acquire for no money.
Which is why I'm on this list.
But I am. One little nook that seems to have been overlooked by the
mainstream computer industry is Synthesis OS, written back in the early 90s.
It could run Unix software *faster* than Unix, on the same hardware [1],
even though it was the furthest thing from Unix you could imagine. The best
explanation is "Just In Time compilation of the kernel" [2], which is even
more impressive when you consider it was written in assembly.
The thesis describing it [3] is mindblowing, and I keep reminding myself
to write the author to see if I could get a copy of the source code, just to
see how it worked.
But while JITs have become popular for languages (and by extension, user
code), no one seems to have picked it up for operating systems, which is a
shame since it seems the concept is just ripe for solving the C10M problem
[4].
-spc (And by "run Unix software faster than Unix," I mean, by an order of
magnitude faster ... )
[1] 68030 based systems.
[2]
http://c2.com/cgi/wiki?SynthesisOs
[3]
http://valerieaurora.org/synthesis/SynthesisOS/
[4]
http://c10m.robertgraham.com/p/manifesto.html