[ih] RAND Unix Port code
jack at 3kitty.org
Mon Feb 13 12:33:19 PST 2017
OK, OK, I'll get my old listing of the Unix TCP and start scanning.
This will generate some huge files. It's probably about 100 8.5x11
pages of stuff. Has anybody figured out the "right" way to do this to
maximize usability (in case someone really wants to OCR or whatever)?
600DPI TIFs? One-file-per-page? And where on the net do you put such
piles of files so they're accessible?
Re: that Unix TCP. My TCP was derived from Jim Mathis' TCP that was
part of the TIU and other boxes built from LSI-11s. Jim's TCP was
written to run on top of MOS (IIRC Micro Operating System). MOS was the
underpinning for the TIU as well. I essentially took Jim's TCP, mixed
well with Unix V6 and Rand ports, reworked to use Unix system services
in place of MOS, reworked Rand ports to be faster, defined await/capac
primitives in Unix, and defined the basic open/close/read/write/etc
Unix-style interfaces for higher level applications to use TCP.
Re Rand ports: I added a few lines to the basic Rand ports code to
improve performance by avoiding writing all data to the underlying disk
file (which the base code actually did although it wasn't obvious;
perhaps unintentional). So you may find some of my fingerprints in
The other major player in the MOS family tree was the Port Expander. A
common configuration was to put a Port Expander between your old
friendly NCP host and its ARPANET IMP. That configuration provided 3
additional ARPANET TCP-only ports, err, sockets, err, you know, the
actual hardware connectors you plugged real physical cables into.
We never did get terminology nailed down....
On 02/13/2017 09:09 AM, Noel Chiappa wrote:
> I'm not sure if this news has spread widely yet, but I have just recovered a
> complete copy of the filesystem for the MIT-CSR Unix (the V6 Unix system at
> MIT-LCS on which most of the early TCP/IP work at MIT was done). I have found
> many treasures therein, including several early TCP/IP's. (More on this
> I'm trying to make them all accessible through Unix source archives:
> for public access. To do so, I need to make sure they are all OK to release
> publicly. (Back in the day, they were not, but that was because the underlying
> UNIX code was protected.)
> One of them is the BBN TCP/IP done by Mike Wingfield. It uses the RAND Port
> code, and I'm trying to work out who can tell me, or OK, the release of that
> Alas, Steven Zucker, the person who wrote the RAND report detailing the
> implementation of ports (and likely the actual author of the code) is no
> longer with us. Carl Sunshine wrote the overview document, but I don't know
> how to reach him - does anyone here? Does anyone else have any information
> on whether or not we can make this code public? Thanks!
> We do not, alas, have the source for the first BBN Unix TCP/IP - the one done
> by Jack Havery as, I am under the impression, a port of Jim Mathis' TIU code.
> We do, however, have the complete TIU source, and I hope to make that
> available too. Same question(s) about that: does anyone know how to reach Jim,
> or know anything about the releaseability of that code? (I seem to recall it
> was done under contract to DARPA, and so probably was open, but not all code
> written under a contract with the USG is necessarily public.)
> internet-history mailing list
> internet-history at postel.org
> Contact list-owner at postel.org for assistance.
More information about the Internet-history