Recommendations for a PDP-10 newbie?

[David Conners]

Thinking of investing some time playing with a PDP-10 image on my spare Pi in advance of the PiDP-10 kit launching.

Because I actually used a PDP-11 with BSD back in the day that was my natural choice for the PiDP-11, but never having used a PDP-10 (that I can remember) I don't know which of the multiple OS choices I should be looking at. And before you say "why not all?", I have small children and a job, so time is limited :-)

Any recommendations on an image or OS to invest in?

Cheers

David

[Johnny Billquist]

I would say TOPS-20 is the nice system with all bells and whistles.
However, no PDP-10 OS is going to lend itself to a quick learn, and Unix
isn't really all that helpful knowledge to have.

Just to warn you a little bit... :-)

Johnny

> --
> You received this message because you are subscribed to the Google
> Groups "[PiDP-11]" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an email to pidp-11+u...@googlegroups.com
> <mailto:pidp-11+u...@googlegroups.com>.
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/pidp-11/44e162d2-471b-41a1-b5e8-c58be94c8cben%40googlegroups.com <https://groups.google.com/d/msgid/pidp-11/44e162d2-471b-41a1-b5e8-c58be94c8cben%40googlegroups.com?utm_medium=email&utm_source=footer>.

--
Johnny Billquist || "I'm on a bus
|| on a psychedelic trip
email: b...@softjar.se || Reading murder books
pdp is alive! || tryin' to stay hip" - B. Idol

[Johnny Billquist]

Um. And I should have added "...as system manager". If you are just a
simple user, then TOPS-20 is actually very easy to live with.

Much better than most anything else, I'd say.

Johnny

[Adam Thornton]

On Sat, Jun 24, 2023 at 9:23 AM Johnny Billquist <b...@softjar.se> wrote:

Um. And I should have added "...as system manager". If you are just a
simple user, then TOPS-20 is actually very easy to live with.

Much better than most anything else, I'd say.

Agreed, and further:

If you're coming from Unix/Linux familiarity, the TOPS-20 Pandas distribution by the late Mark Crispin is excellent and easy to get started with.  That plus https://www.math.utah.edu/~beebe/publications/1987/t20unix.pdf make it such that I, a complete newbie, was able to port Frotz to TOPS-20 without very much friction from the operating system (the presumption in the source code that bytes are 8 bits was more difficult).

Adam

[David Conners]

Much obliged for the recommendation.

I'm absolutely expecting a learning curve, but can't justify getting that gorgeous reproduction Oscar has created if it is just going to function as a fancy case for a Raspberry Pi ;-)

Cheers

David

[Ken Hansen]

I'd be interested in pointers to setting up an RPi to emulate the PDP-10, also with the goal of getting a jump on learning the PiDP-10.

I remember the late Paul Allen put a PDP-10 on the internet, anyone have any info on that effort? Did the system survive him, or has it gone 'off the air' as they say?

Ken

On Jun 24, 2023, at 11:40, David Conners <da...@conners.com> wrote:

Much obliged for the recommendation.

--

You received this message because you are subscribed to the Google Groups "[PiDP-11]" group.

To unsubscribe from this group and stop receiving emails from it, send an email to pidp-11+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/pidp-11/9679611f-0cc3-4ac2-a30a-80401582f519n%40googlegroups.com.

[Johnny Billquist]

Tricky thing, though. It's supposed to be a KA10. And that CPU can not
run TOPS-20. Not sure what can be cooked together to get around that...

Johnny

> --
> You received this message because you are subscribed to the Google
> Groups "[PiDP-11]" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an email to pidp-11+u...@googlegroups.com
> <mailto:pidp-11+u...@googlegroups.com>.
> To view this discussion on the web visit

> https://groups.google.com/d/msgid/pidp-11/9679611f-0cc3-4ac2-a30a-80401582f519n%40googlegroups.com <https://groups.google.com/d/msgid/pidp-11/9679611f-0cc3-4ac2-a30a-80401582f519n%40googlegroups.com?utm_medium=email&utm_source=footer>.

[Clem Cole]

There are a number, and each has advantages, although coming from UNIX TOPS-20 (TENEX) are likely to be the easiest.  The biggest lift is likely to be the editor if you don't EMACS already (more in a minute).

Start here with:

TOPS-20 SOFTWARE NOTEBOOKS - which is a list of all the DEC documents

TOPS-20 Command Reference

UNIX for TOPS-20 Users - Nelson's Beebe [who is still quite active on the TUHS mailing list] wrote a great doc to go the other way back in the day.

There are several C compilers around. In fact, the first C outside of the PDP-11 is the Synder compiler from the 1970-2 timeframe.  But .... there are dragons here because chars are 7 bits (5 chars to a 36-bit word), which was how most HLLs in the 10/20s worked back in the day.  So choose the compiler wisely and understand what you are getting [I'm not sure what is in the Panda distribution, so I can not say].   FWIW: Old V5/V6 PDP-11 C code often masks off the top bit due to that how the TTY driver worked, so moving UNIX tools from the old systems using the Synder compiler was not awful because much of the code only needed 7 bits.    But if you are moving more modern code, you may need to find a more modern C implementation - which I suspect there are a few to choose from -- I have not looked in at least 5-10 years (Lars Brinkoff is probably the best person to ask about PDP-10 tools - of course, he likely to tell you to run ITS, not TENEX 😎

BTW: Lar's is wonderful, generous, and kind and has a great deal in his libraries of different PDP-10 lore.

Editors -- the closest thing to UNIX ed(1) on the 10s and 20s out of the box was/is SOS - Son of Stopgap.  You can probably use it to get started if you know ed.   Actually, I went the other way from SOS to ed back in the day.   Teco and EMACS were the primary editors, if you want to say true to the times ... however ...

• Ward Millers - s editor from his SW tools book (https://github.com/udo-munk/s) compiled on the 10 family a long time ago.  That's a subset vi for the VT-100 might be a good place to start - its small and simple and if you know the basic vi/ed commands - it will be easy to edit things.
• Many of the UNIX tools were ported/written for the 10s and are in places like the DECUS library.
• I saw a vi implementation in the early 1980s, but I never used or had it when using the PDP-10/20s in the mid-1970s.  Nelson Bebe might know more about it, since Utah was a big UNIX and TOPS-20 place for a number of years.  [Back in the days, there were vi ports for everything, the Cray, CP/M, DOS, etc. - which I did/do have squirreled away].
• EMACS is written for ITS (as a set of TECO macros) and moved to TENEX, TOPS-10, and WAITS.   Gosling reimplemented in C for VMS/UNIX in the late 1970s/early 1980s [which forked later to become GNU].
• TECO was more popular at DEC in those days (and that's what I used) -- but I'm not sure I could recommend it.  As it has been said, TECO command resemble "line noise" from when we had dial-up modems.

Good Luck,

Clem

ᐧ

--

You received this message because you are subscribed to the Google Groups "[PiDP-11]" group.

To unsubscribe from this group and stop receiving emails from it, send an email to pidp-11+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/pidp-11/44e162d2-471b-41a1-b5e8-c58be94c8cben%40googlegroups.com.

[Clem Cole]

below...

On Sat, Jun 24, 2023 at 1:28 PM Ken Hansen <n2...@w5fc.org> wrote:

I'd be interested in pointers to setting up an RPi to emulate the PDP-10,

OpenSIMH.org -- This is the emulator used by Oscar - you can just build the PDP-10 variants.  There are doc and examples on the website.

Mark Crispen's Panda distribution: Panda TOPS-20 distribution

 

also with the goal of getting a jump on learning the PiDP-10.

I remember the late Paul Allen put a PDP-10 on the internet, anyone have any info on that effort? Did the system survive him, or has it gone 'off the air' as they say?

Sort of ...   the museum is closed thanks to CV-19:  LIVING COMPUTERS: MUSEUM + LABS IS CLOSING FOR NOW

Some of the systems were still available if you had accounts on them, although I'm not sure which ones are still alive.

ᐧ

ᐧ

[Warner Losh]

On Sat, Jun 24, 2023 at 11:55 AM Clem Cole <cl...@ccc.com> wrote:

But .... there are dragons here because chars are 7 bits (5 chars to a 36-bit word), which was how most HLLs in the 10/20s worked back in the day.  So choose the compiler wisely and understand what you are getting [I'm not sure what is in the Panda distribution, so I can not say].

There's fine-grained control over the character size in many of the compilers (beyond C too), so be careful they all agree on the character size and packing conventions... This was a problem back in school when we had to make a C, Pascal, fortran, macro-20 (and maybe an exotic like simula or clu) program for a language survey class I had (one program, all those languages).

 

TECO was more popular at DEC in those days (and that's what I used) -- but I'm not sure I could recommend it.  As it has been said, TECO command resemble "line noise" from when we had dial-up modems.

TECO also had a visual mode that wasn't horrible...  But most people eschewed that and used either the direct commands, which were kinda like ed, but also very much not at all like ed (same neighborhood as ed relative to what most modern users are used to, but on the other side of the planet where the water circles the drain differently). TECO is fun to play with and does some simple things easier than ed, but its lack of regular expression support is limiting.

One thing not mentioned: The extensive command line completion. Makes bash, tcsh, and zsh look primitive since there was a hook directly into all programs that parsed command line arguments that is lacking in Unix (or some might say that Unix wisely avoided, it all depends on your philosophy). It's really nice to be able to hit escape, or ? and get instant help for what /glortz means (or really it should be /glitz).

Warner

 

Good Luck,

Clem

ᐧ

On Sat, Jun 24, 2023 at 12:11 PM David Conners <da...@conners.com> wrote:

Thinking of investing some time playing with a PDP-10 image on my spare Pi in advance of the PiDP-10 kit launching.

Because I actually used a PDP-11 with BSD back in the day that was my natural choice for the PiDP-11, but never having used a PDP-10 (that I can remember) I don't know which of the multiple OS choices I should be looking at. And before you say "why not all?", I have small children and a job, so time is limited :-)

Any recommendations on an image or OS to invest in?

Cheers

David

--
You received this message because you are subscribed to the Google Groups "[PiDP-11]" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pidp-11+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/pidp-11/44e162d2-471b-41a1-b5e8-c58be94c8cben%40googlegroups.com.

--
You received this message because you are subscribed to the Google Groups "[PiDP-11]" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pidp-11+u...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/pidp-11/CAC20D2PWtbEqhvVdg6OHfzG4X55YLq%3DqEPz1uEcQhKEvMD%2Bv9g%40mail.gmail.com.

[Clem Cole]

BTW, I just chatted with Mr. PDP-10 this PM about KA/KI/KS/KL cpu and DEC/BBN/MIT MMU/pagers -- WRT to SIMH Simulators vs. Ken Harrenstien's KLH10 emulator (KLH10 PDP-10 Emulator)

While we have been suggesting you learn to start/run TOPS-20 (which I believe both will support), it is not going to be accurate on Oscar's new kit, which is a KA10 front panel.  So if you plan to get ready to run it as an accurate emulation, you'll need to run TOPS-10 or ITS with the code in Supnik's SIMH [he has a command: set cpu ITS MPX 1024K in the KA emulator].   But I don't think the BBN pager is emulated, which is an issue.

TENEX was written for the BBN pager (which is different from the MIT pager or KS or KL), but there were versions of it working on different PDP-10 implementations around the ARPANET before DEC picked it up to create TOPS-20.   However .... as I verified from Tim:

TOPS20 would never run on a KA.  That code was ripped out early on.  I don't think there was an attempt to run it (or TENEX) on the KI; the VM subsystem depended on the core status/age features.  Hardwired TTL logic; might have been possible to emulate, but trapping every write ref would have been slower than a KA.

T20 was tightly married to the KL from day 2.
T20 has some remnants of support for some of the KA IO bus peripherals (cards and printers, and the IMP), but depends heavily on the native RH20 massbus registers for paging IO.  All disk IO in t20 is paging.
T20 on the KS does work, but is crippled by the Unibus IO, even with some special hacks on a dedicated unibus/RH11 for disk.

Hopefully, this is helpful.

Clem 

ᐧ

[Heinz-Bernd Eggenstein]

Note there is a PiDP-10 google group now as well. Perhaps that is a better place to discuss this stuff.

[David Conners]

Thanks all for the information, have requested to join the PiDP-10 group and will research the posts there.

Cheers

David

[steve...@gmail.com]

Just a quick FYI, but I believe TOPS-10 was the dominant OS on KA-10 CPUs.

Anticipating Oscar's KA-10 I've been playing around with TOPS-10 on the Blinkenbone KL-10 simulator:

http://retrocmp.com/projects/blinkenbone/simulated-panels

-- steve

[Lars Brinkhoff]

Some minor remarks.

- The PiDP-10 will probably support TOPS-20 V7 (running on a KL10 model B), somehow.  We don't know exactly how yet.

- EMACS does not run on TOPS-10 or WAITS.  WAITS has the excellent E editor.

- Cornwell's PDP-10 emulator supports the BBN pager.  It's used by WAITS, but we haven't tried TENEX yet.

- TENEX did run on the KI10, using the KI10 paging hardware instead of the BBN pager.  This version has shown some sight of life on the emulator.

[Johnny Billquist]

A remark on the remark then...

On 2023-06-27 13:02, Lars Brinkhoff wrote:
> Some minor remarks.
>
> - The PiDP-10 will probably support TOPS-20 V7 (running on a KL10 model
> B), somehow.  We don't know exactly how yet.

That will be interesting. Obviously simh emulating a KL is desireable.
Not sure if it's there yet, but I expect it to be at some point. And the
PiDP-10 would probably be running simh. The interesting thing is in
which way you make the front panel work with a KL. This would clearly
not be close to anything you would ever have seen in real life, but it's
just that this would be sortof what you want to run.

> - EMACS does not run on TOPS-10 or WAITS.  WAITS has the excellentE editor.

There is the very nice Emacs clone called AMIS that is available for
Tops-10. Done by some nice T10 hackers in Sweden back in the 80s. I
would expect that you know about this.

Johnny

[Lars Brinkhoff]

Johnny Billquist wrote:
> That will be interesting. Obviously simh emulating a KL is desireable.
> Not sure if it's there yet, but I expect it to be at some point.

Cornwell updated his emulator to do both KL model A and B a while ago.
(And, lastly, KS too.)

> The interesting thing is in which way you make the front panel work
> with a KL.

It would probably be blinking prettily in some way, but certainly not
realistically.
Perhaps not even useful. Note that MIT-MC did have an add-on board to
provide blinking lights, so in some way there's a precedent...

> There is the very nice Emacs clone called AMIS that is available for Tops-10.

Right, and there's another one called FINE.

[Heinz-Bernd Eggenstein]

While I'm waiting for Oscar to let me into the PiDP-10 group (no pressure...), maybe I can ask a question here:

I think I want to explore ITS on the KA10  from here : https://github.com/PDP-10/its  (which looks fascinating), but I run into timeouts (a known problem it seems) when trying to build this on a RaspberryPi like the one I'll use for the PiDP-10. What takes so much time, even on a relatively fast Linux PC, are the steps that build tapes and disk images from within the emulator.

Now, would it be possible to transplant the tape and disk images produced on a 64 bit Linux PC and put them into the repsective parts of the build tree of the (32 bit OS) PI and let the build process pick up from there? Or are some specific bits of information like network address etc baked into the images that would make them unuseable on a different host ?

Thanks in advance

HBE

[Lars Brinkhoff]

HBE wrote:

I think I want to explore ITS on the KA10  from here : https://github.com/PDP-10/its  (which looks fascinating), but I run into timeouts (a known problem it seems) when trying to build this on a RaspberryPi like the one I'll use for the PiDP-10.

That's right.  The build script is not tuned to run on a Raspberry Pi, although some people have been working on that recently.  The culprit is mostly slow access to the SD card file system.  I have a Pi 4 with my home directory on an USB memory stick, and that is fast enough.

 

Now, would it be possible to transplant the tape and disk images produced on a 64 bit Linux PC and put them into the repsective parts of the build tree of the (32 bit OS) PI and let the build process pick up from there?

Yes, that is entirely possibly.  The disk images are portable across any type of host.  There is already a "make download" target that supposedly downloads the latest disk images and places them in the right spot.  This is somewhat lightly tested, so please report if it doesn't work.

You can also download the disk images manually from here:

http://hactrn.kostersitz.com/

Disk space courteously hosted by former ITS exporer Mike Kostersitz.

[Heinz-Bernd Eggenstein]

Excellent, thanks for the hints!! The "make download" is a really good idea, and with it I was finally able to complete the build on a Rasberry Pi 3 (still on Buster). Next things is reading some instructions and try to start up the sim and make sense of it.

I really hope that the PiDP-10 panel will blink happily with ITS. With all the recent hype around AI, it would be facinating to step back in history and go back to the MIT AI Lab of Minsky et al where many of the foundations were laid.

Thanks  again

HBE