| PDP-11 stories (1) [message #400091] |
Wed, 16 September 2020 09:29 |
|
|
Originally posted by: Bob Eager
I was inspired, yesterday, to gather some PDP-11 stories. I hope they are
vaguely interesting.
Our first PDP-11 (in a university) was purchased as a front end to our
ICL 4130 mainframe; it was a cheaper way of getting an extra terminal
multiplexer than buying one from ICL. It worked well, with a locally
developed interface (two colleagues developed one end each).
It worked really well, and inspired the department (both academic and
'service' computing, in the Cambridge model) to buy many more PDP-11s.
The next one was again to augment the ICL 4130. It used the same kind of
interface, and this time it was a 'file server'. It had two RP02 disks
(so about 40MB total), and given that this amount of storage would have
required five more disk controllers and 18 more disks on the 4130, it was
a cheap deal. It used a complicated mapping between usernames and
filenames, the PDP-11 running under DOS-11 and its file system.
I broke it once. I was experimenting with a new program in 4130 assembler
(anyone remember NEAT?). I managed to get it into a loop writing to disk,
and wrote a *very* large file to the PDP-11.
That was OK until I then deleted it. DOS-11 used a fairly inefficient
block chaining technique for files (a common one in those days). So it
had to chain down the file, resetting the allocation bitmap for each
block in turn. This took a *long* time.
Too long. The PDP-11 went deaf while this was happeninh. The 4130 timed
out the link, and did the only thing it had been told to do in that
situation - crash.
--
Using UNIX since v6 (1975)...
Use the BIG mirror service in the UK:
http://www.mirrorservice.org
|
|
|
|
Members
Search
Help
Register
Login
Home
