I'm looking into the history of Spider and
early Datakit. Sandy Fraser
was kind enough to send me the 1974 report on Spider
Is that online? If not, any chances you can make it so?
It is a paper copy, but I can make a scan for you.
which contains the drivers tiu.c, mpx.c - I'm not
sure what other files there
are part of it?
I think tiu.c might be all. The TIU ("terminal access unit") was the network
card,
so to speak (actually some 5 boards in a rack) and did a lot of the heavy lifting.
From the tiu.c file I understand that a DR11-B parallel I/O card was used on
the PDP side to connect to the TIU, and that access was structured as a block
device driver.
I'm not at all clear how this stuff got there -
someone at Bell must have just
dumped the contents of the 'dmr' directory, and sent it all off?
Looks like it.
The PWB1-based MIT systems also have a lot of the
Spider software (although it
was never used). It's a slightly different version than the one above:
'diff'
shows that 'tiu.c' is almost identical, but mpx.c has more significant
differences.
It also contains man pages, and sources for some (?) user programs; I have the
source and manpage for 'nfs'. What other names should I be looking for? (The
man page for 'nfs' doesn't list any other commands.) I'll put them
up
momentarily.
I think that in the lifespan of Spider (1972-1978) there were 3 main network
programs (basing myself on McIlroy's Unix Reader):
- 'nfs' an FTP-like program to copy files to/from a central File Store.
I'm not sure whether the File Store was a Unix machine or something else.
- 'ufs' not sure what it was, but I think a telnet-like facility
- 'npr' a network printing program
A little surprising, but no reference to a Spider mail program in that document.
In the meantime, I'll append the 'tiu'
man page.
Thanks! It is from October 1973, which sounds right for Spider. I guess this
code is the first networking on Unix, predating the UoI work by about 18 months.