V8's roots?

[norman at nose.cita.utoronto.ca]

The operating system kernel on the V8 distribution tape (which was sent
to less than a dozen places, under special license) was indeed descended
from one of the 4.1 BSD releases; I have a vague memory that it was 4.1a,
but I wasn't there at the time, and don't know just what was in each of
the intermediate 4.1s.  As I understand the history (again, I wasn't there
when this part happened), when the Computing Science Research Center
decided to move its main computing world to VAX in the early 1980s,
they wanted a reasonably stable, reasonably fast system with paging,
and 4.1x (for whatever value of x it was) seemed the best available choice.
The only real competitor was the paging descendant of 32/V done by John
Reiser (who did the original 32/V port to the VAX, I believe), but that
system seemed to have lost the evolutionary battle and was judged a bad
bet.

It may help to identify the kernel in question to know that it probably
didn't have sockets yet, and certainly didn't have FFS.

The 4.1x kernel was just used as a base, however.  By the time I arrived
at the Center in late 1984, a good bit had been added and replaced: the
V7-heritage terminal IO subsystem had been kicked out in favour of Dennis
Ritchie's stream I/O system; Peter Weinberger's simple disk file system
speedups (4KB blocks and a bitmapped free list, nothing more) and network
file system code and the corresponding file system switch had been added;
Tom Killian's process file system had appeared.

The commands in /bin and /usr/bin and whatnot had less obvious BSD influence,
and I suspect they were mostly carried over from the system used internally
on the PDP11s when the VAXes first arrived.

Norman Wilson
(six years in New Jersey drove me out of the country)

Received: (from major at localhost)
	by minnie.cs.adfa.edu.au (8.9.1/8.9.1) id KAA15731
	for pups-liszt; Thu, 7 Jan 1999 10:29:31 +1100 (EST)