Join me, together we can rule the ARPANET

[Lars Brinkhoff]

Hello,

I'm working on setting up an ARPANET using emulated IMPs with an interface to talk to

emulated hosts.  Since the selection of hosts with available NCP software is mostly

limited to PDP-10 systems, I turn to you fellow PiDP-10 enthusiasts.  I would love to

see a distributed ARPANET with some of the original hosts and many a few new ones.

I call dibs on IMP 6 and maybe 11.  Which one do you want?

[Paul Birkel]

Lars;  I’m not sure of the specific IMP numbering assignments.  Is https://en.wikipedia.org/wiki/ARPANET#/media/File:Arpanet_map_1973.jpg too late for your purposes?  I have a fond spot for CARNEGIE there, although I didn’t reach CMU until Fall 1981 and VAX’en were ascendant by that point … I did manage to get time on a PDP-10 for a while.  I’ve been looking forward to the PiDP-10 as a “way back” means to revisit that period of my youth.  Would be pleased to host the CARNEGIE IMP, if that works for you/others.  Not that I have much of a clue what would be involved, yet ;->.

--
You received this message because you are subscribed to the Google Groups "PiDP-10" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pidp-10+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/pidp-10/9fde338f-6c46-4e88-9c4e-adcf4b806a44n%40googlegroups.com.

[Lars Brinkhoff]

Hello Paul,

That would be IMP 14 (octal).

1973 is actually pretty spot on if you consider the few copies of IMP software

we still have.  See https://walden-family.com/impcode/ for the full story.  But

I think we can tolerate some anachronisms.

The only requirement is to have a computer online (real computer on the real

Internet) wich a few UDP ports open.  Those are your modem lines to other

IMPs.  On that computer you run the SIMH H316 emulator.  What you attach

to it is entirely up to you, just like back in the day.  For CMU that would preferably

be TOPS-10 I guess.  Let's hope some NCP software can be found.  It's also

possible to write something new in, say Python or something, if you want to

go wild or just mimic something cool that now has been lost.

[James Bond]

I think Lars is correct about just TOPS-10 being on the ARPANET.  Though CMU had DEC-20s running TOPS-20, AFAIK they weren't on the ARPANET by the time I graduated in 81.  I have a vague feeling that some of the PDP-11s that the CS department owned had some flavor of ability to connect the ARPANET, but that's definitely not certain.  I do know that we were able to telnet and NFT files from MIT-XX and MAXC back in the day; as well as telnet the other MIT- machines (AI, etc) via the ARPANET in say 1978/79.

Steve

To view this discussion on the web visit https://groups.google.com/d/msgid/pidp-10/124201d7c680%24d684cd20%24838e6760%24%40gmail.com.

[Lars Brinkhoff]

I  think Lars is correct about just TOPS-10 being on the ARPANET.  Though CMU had DEC-20s running TOPS-20, AFAIK they weren't on the ARPANET by the time I graduated in 81.  I have a vague feeling that some of the PDP-11s that the CS department owned had some flavor of ability to connect the ARPANET, but that's definitely not certain.

To answer these kinds of questions, I usually go to https://github.com/ttkzw/hosts.txt/blob/master/pub/hosts/19830527/HOSTS.TXT

We can see a CMU-20C in 1982, so very much at the tail end of the NCP era.  Before that it seemed a TOPS-10 shop, and of course Hydra

running on C.mmp!  There's an emulator I'd like to see.

[Andy]

Hi Lars - please sign me up as an ARPANET host!

I am not an expert at all in this field even slightly... but I can certainly follow written instructions.... and I do just happen to have two Pis running TOPS10 right now  - are these of any use?

cheers!
Andy

[Lars Brinkhoff]

Andy wrote:

Hi Lars - please sign me up as an ARPANET host!

Great!  I guess you TOPS-10 people will have to fight it out amongst yourselves which sites to host.  Say, do you happen to live near Harvard?

 

I do just happen to have two Pis running TOPS10 right now  - are these of any use?

Well, first of all I did write "I'm working on setting up an ARPANET", by which I mean it's not ready yet.  Actually, I just got a host-to-host packet roundtrip to work only today.

Second, being an ARPANET host isn't all roses.  You should have an NCP, which is the system Network Control Program.  Unfortunately, those are not easy to find.  I do not yet know about any NCP for TOPS-10.  An impoverished by workable alternative might be to just emulate remote login through some kind of gateway.  Maybe throw something together in Python or something?  Perl?  TCL?  Pick your poison.

IMP emulates modem lines using UDP packets, and the host interface works the same way.

Or maybe hack the TOPS-10 TCP/IP stack to do NCP instead?

[Lars Brinkhoff]

Oscar picked up the project and made this:

https://obsolescence.dev/arpanet_home.html

We have the three MIT ITS hosts online.  SAIL is also there, but in a fake way because I'm not sure the networking code is ok yet.  Maybe BBN's TENEX machines will be next.

Here are some things to work on:

- Illinois' PDP-11 running Unix.  We have the code, but there are gaps that need to be filled.

- ELF on the PDP-11.  We have the code and it assembles fine, but we need a linker.

- I have a lead on the Lawrence Livermore lab PDP-11 running RISOS.  Fingers crossed.

- The TOPS-10 NCP is lost (or in hiding).  Maybe the one from SAIL can be ported over?

- The UCLA 360 running OLS, is there any software left?

- Other IBM mainframes, what to do?

[Charley Jones]

Go, Lars, Go!

Sent from my iPhone 15pm!

Charley Jones, PMP

On Jan 22, 2026, at 10:17 PM, Lars Brinkhoff <lars.br...@gmail.com> wrote:

Oscar picked up the project and made this:

--

You received this message because you are subscribed to the Google Groups "PiDP-10" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pidp-10+u...@googlegroups.com.

To view this discussion visit https://groups.google.com/d/msgid/pidp-10/5091e8e8-8f73-4fb7-9e3b-224c18be1416n%40googlegroups.com.

[Mitchell Wolrich]

Lars,

MIT had a fourth ITS PDP-10, “MIT-MC”, when “Macsyma” split of MIT-ML and was hosted on a dedicated PDP-10..  I was in High School in the mid 70’s, and used to “hang out” at 545 Tech Square, 9th floor, which had 4 PDP-10’s, and the infamous “Star Wars” PDP-1.  it was my first exposure to news groups, email, and what would become the internet 👍😂

Mitch

Sent from my iPad

On Jan 23, 2026, at 1:17 AM, Lars Brinkhoff <lars.br...@gmail.com> wrote:

Oscar picked up the project and made this:

--

[Lars Brinkhoff]

Mitchell Wolrich wrote:

MIT had a fourth ITS PDP-10, “MIT-MC”, when “Macsyma” split of MIT-ML and was hosted on a dedicated PDP-10.

Quite so.  However, Oscar's network is aimed at recreating the ARPANET as it was around 1972-73, and at that point there was no MIT-MC.

I was in High School in the mid 70’s, and used to “hang out” at 545 Tech Square, 9th floor, which had 4 PDP-10’s, and the infamous “Star Wars” PDP-1.

Please tell us more about what you saw up there!   The Spacewar computer must have been the ailing AI lab PDP-6.  The game was first written for a PDP-1, but it was ported to the 6.  The AI lab did once have a PDP-1 and traded it in for a 6.  Here is a series of photos of the AI PDP-6.  http://its.pdp10.se/pdp6-timeline/

There were two more PDP-6'es around Tech Square: the MIT Dynamic Modeling group had one as well, as did Keydata (not Tech Sq 545 though).

[Mitchell Wolrich]

yep, it was probably the PDP-6, I remember the Space War funky controllers, built out of small wooden boxes, played a few times, it was difficult..  it had simulated gravity that sucked everything into the center (could have been a blackhole).

back in the mid-70’s, MIT had a program called “HSSP” (High School Studies Program), where we could take “morning” and “afternoon” college level classes, in unused MIT classrooms, usually taught by MIT grad students.  between sessions, a few of us would wander over to 545,  we met a few early “hackers”..  one of who, actually created acounts for us on the MIT-AI PDP-10, my first introduction to ITS & how I signed up on “list servers” for those early newsgroup emails.  of course some of the MIT professors were not thrilled with a bunch of high school kids on the 8th & 9th floors, we where eventually told not to visit 😳   by then, I had the phone # for an early BBN dial up, to a terminal server, so I could still check email, remotley via a 300 baud modem on a dumb terminal that a friend of my parents who was a VIP at Prime Computer lent me, eventually, I built an Altair 8800 as well, and had one of the first CBBS systems in Boston, it ran under CP/M on (2) old 8” floppy drives, and had an improved (up to 600 baud) dial up modem 😂  it was featured on the cover of January 1980 Kilobaud Magazine (where I eventually worked for 2 summers while in High School)

FYI, the MIT HSSP program still exists today 👍

Mitch

Sent from my iPad

On Jan 23, 2026, at 5:35 AM, Lars Brinkhoff <lars.br...@gmail.com> wrote:



[Andy]

Hey Lars - whats the status of collecting IBM ... any ideas?

Here are some things to work on:

- Illinois' PDP-11 running Unix.  We have the code, but there are gaps that need to be filled.

- ELF on the PDP-11.  We have the code and it assembles fine, but we need a linker.

- I have a lead on the Lawrence Livermore lab PDP-11 running RISOS.  Fingers crossed.

- The TOPS-10 NCP is lost (or in hiding).  Maybe the one from SAIL can be ported over?

- The UCLA 360 running OLS, is there any software left?

- Other IBM mainframes, what to do?

[Lars Brinkhoff]

Andy wrote:

Hey Lars - whats the status of collecting IBM ... any ideas?

Not much.  One could always have a networking frontend which directs login sessions to the mainframe.  But that would mean telnet only, and no outgoing connections from the mainframe. 

[Noah Smith]

Hi Lars, we'd like to volunteer to adopt the utexas node as part DecwarOrg - suspect utexas was a bit after 1973 though, maybe around 1975 or so as a guess? Will start digging into that topic!:) Please pencil us in for utexas (where it makes sense for you:), and we're linking into all this already from our new website, down in the 'resources' section:) 

https://decwar.org/

Cheers,
Noah

[Noah Smith]

fwiw, this below looks about right to me - so decwar.org node is a 'few years late':) but we'd love to be part of the story in whatever role!:)

1. Confusion with the "First Four" (1969)

The "UT" in the original four ARPANET nodes refers to the University of Utah (Node 4), which was installed in December 1969. The other three originals were UCLA, SRI (Stanford Research Institute), and UC Santa Barbara.

2. Absence in Early Maps (1969–1974)

• 1969–1973: UT Austin does not appear on ARPANET logical maps from this period.

• January 1974: A management study report from January 1974 explicitly noted that major commercial and research centers in "Texas are not" yet represented on the network, confirming that UT Austin was not yet connected at this time.

3. Installation and Connection (1975–1976)

• Appearance on Maps: The University of Texas (often labeled "Texas" or "UTEXAS" on logical maps) begins appearing on ARPANET maps between 1975 and 1977.

• Visual Confirmation: By the March 1977 logical map, "Texas" is clearly visible as a node, connected to the network alongside other expanding universities.

• Directory Evidence: The ARPANET Directory from 1978 lists the University of Texas at Austin (specifically the Linguistics Research Center and Computation Center) as a fully operational host.

[Lars Brinkhoff]

Noah Smith wrote:

Hi Lars, we'd like to volunteer to adopt the utexas node as part DecwarOrg - suspect utexas was a bit after 1973 though, maybe around 1975 or so as a guess?

There's room for more than one ARPANET project.  Oscar is mainly aiming for an around 1973 network with historical hosts, and run them all on a single machine.  I my first idea, and how I started this thread, is to have a somewhat later vintage network operated by enthusiasts connecting their computers together.

An question for you is whether UTEXAS ran TOPS-10 or TOPS-20?  I have some vague impression DECWAR was more of a TOPS-10 thing, but maybe I have that wrong.  If UTEXAS was a TOPS-10 shop, the bad news is that we haven't located an NCP for that operating system.

[Noah Smith]

Sounds good Lars, understood, and we'll keep track of what's happening and do what we can for UTEXAS node where it makes sense. Definitely TOPS-10 for DECWAR unfortunately. Dual KL-10. And there was a dual CDC 6600/6400, the 'Dual Dinosaurs' as they were called. Those were the biggies there in the Computation Center along with the IMP or TIP.

Cheers, Noah  

[Lawrence Fisher]

Not sure of the iMP address, but I would enjoy reserving USC-ISI's imp!

[Lars Brinkhoff]

Is anyone interested in having TOPS-20 on the network?  I looked around, and the chance of success seems good.  I jotted down some pointers here:

https://github.com/obsolescence/arpanet/wiki/TOPS%E2%80%9020AN