Wednesday, February 10, 2010

ChaosNET implementations for PDP-10

This is not particularly interesting, but I wanted to record a couple of links to ChaosNET code for the PDP-10. The ChaosNET memo describes an ITS (Incompatible Time-Sharing System) interface, for which the code is somewhere in the ITS sources, and a Lisp Machine implementation, for which code is in the CADR sources and Symbolics code base, but also a TOPS-20/TENEX implementation, a VMS/VAX implementation, and a UNIX implementation. The UNIX implementation was documented by James E. Kulp for Symbolics, and the code seems to have survived. The TOPS-20/TENEX is harder to find, but it looks like BitSavers' Foonly archives have some files.


I have no idea whether this is at all related to the implementation the AI MEMO mentions.

A Google search turns up some interesting code (might be for the PDP-11 front end, for ITS)


CH10.defs1 "PDP-10 chaosnet interface definitions"
CH11.defs1 "CH11 Unibus chaosnet interface definitions"
CHSDEF.15 "Chaos net packet definition (for PDP-10s)"
KSNET.1 "Chaosnet interface on Unibus" (presumably for the KS10)
IOELEV.432 "i/o eleven"
NET.32 General ITS networking code


  1. The ITS files mentioned are written in MIDAS, the MIT assembler for the PDP-10. They are part of the ITS source code. As far as I know, the only ITS system with a PDP-11 front end (MIT-MC) did not have a Chaosnet interface on it. The references to a Unibus in the ITS code guarantee that it was written for the KS-10 processor ("DECSYSTEM-2020"), which is a PDP-10 implementation which uses the Unibus internally.

    (As it happens, I'm working on restoring the MIT-AI KS-10 system to working condition, including re-installing ITS with more disks available, at the Living Computer Museum in Seattle.)

  2. Thanks for the info. I guess I was keying off of the CH10P and various "11" flags in the code. The AI memo also talks about byte-order issues that made it sound like the PDP-10s on Chaosnet would be using PDP-11 front-ends to connect.

    I don't know much about ITS (or DEC systems), except that it seemed to be the environment in which the early Lisp machines came to life. I'm probably missing enormous amounts of context to really understand how things were. I was using Brad Parker's CADR emulator, and had issues getting his Unix Chaos FILE server to work, and thought I might have better luck bringing up a simulated ITS system for the simulated CADR to talk to.