Kermit and TOPS-10: to Linux works, to TOPS-10 doesn't

[Ric Werme]

One of the first things I wanted to do on PiDP-10's TOPS-10 was transfer an old .MAC file to it. Kermit didn't work, and I resorted to reading from PTR: and writing to LPT:.  Have others run into this problem?

Now, weeks later, I wanted to look into the problem, and found Kermit was works fine to pull files down to Linux but fails immediately on upload.  On the client (Linux Mint desktop, not the PiDP-10's Raspberry Pi) I see:

(~/pdp10/) C-Kermit>put fildmp.mac

*************************
SEND-class command failed.
 Packets sent: 1
 Retransmissions: 0
 Timeouts: 0
 Damaged packets: 0
 Transfer canceled by receiver.
 Receiver's message: "?KERDAT   Transmission not in sync"

HINTS... If the preceding error message does not explain the failure:
 . Try changing the remote Kermit's FLOW-CONTROL setting.
 . Try it again with: SET STREAMING OFF
 . Try it again with: SET PARITY SPACE
 . As a last resort, give a ROBUST command and try again.

Kermit on TOPS-10 seems not to know about streaming, it was in "parity none."  "Parity space" fared the same, I have it running in server mode.  I'll poke around some more and play with simh parameters too.  I did enable some debug output in simh, here's the relevant data from a put attempt.  Nothing looks immediately helpful, but not knowing Kermit's protocol, I'm not surprised.

DC RCV: Line:1 Received 10 bytes 'kermit -r_TN_CR_'
DC RCV: Line:1 Received 28 bytes '_^A_9 S~/ @-#Y3~^>J)0___\"U1AV_TN_CR_'
DC XMT: Line:1 Sent 17 bytes '_^A_- Yp% @-# 1~^8_TN_CR__TN_CR_'
DC RCV: Line:1 Received 16 bytes '_^A_-!FFILDMP.MAC0_TN_CR_'
DC XMT: Line:1 Sent 9 bytes '_^A_#!Y?_TN_CR__\000__\000__TN_CR_'
DC RCV: Line:1 Received 58 bytes '_^A_W"A."U1"#AMJ*!A#120251004 22:11:00!"111%11124,#664-!3@ ;_TN_CR_'
DC XMT: Line:1 Sent 41 bytes '_^A_C"E?KERDAT_^I_Transmission not in sync__TN_CR__TN_CR__TN_CR__TN_LF_'

[Noah Smith]

hi - was using kermit with tops10 fine last year (have other workflows now) -
can remember it took some time to get everything right -
any case - here's the notes from all that in case there's anything helpful:)

https://gitlab.com/decwar/utexas/-/blob/main/docs/sec2-kermit.md

cheers, noah

[steve...@gmail.com]

I would have replied earlier but I was out of town on business. Here's the way I've been using Kermit on Tops-10 7.03. The things I type are highlighted in color:

1) Start with a RPi text file to transfer in /opt/myfiles/myfile.txt


2) Open a terminal window on the RPi, cd into the RPi directory
   containing the file, and run kermit on theRPi:

pi@raspberrypi:~ $ cd /opt/myfiles
pi@raspberrypi:/opt/myfiles $ kermit


C-Kermit 9.0.302 OPEN SOURCE:, 20 Aug 2011, for Linux+SSL+KRB5
 Copyright (C) 1985, 2011,
  Trustees of Columbia University in the City of New York.
Type ? or HELP for help.


3) Telnet into Tops-10 on the PiDP-10 and log into the user account. Note that the public account on this install has no password.

C-Kermit>telnet localhost:2020
 DNS Lookup...  Trying 127.0.0.1...  Reverse DNS Lookup... (OK)
Connecting to host localhost:2020
 Escape character: Ctrl-\ (ASCII 28, FS): enabled
Type the escape character followed by C to get back,
or followed by ? to see other options.
----------------------------------------------------


Connected to the PDP-10 simulator DZ device, line 1


.login public
Job 11  VNW  TTY1
[LGNJSP Other jobs same PPN:7,10]
20:33   27-Aug-21   Friday

4) Run Kermit on Tops-10 and configure it to receive text files

.r kermit

TOPS-10 KERMIT version 3(136)
Kermit-10>set file byte-size seven-bit
Kermit-10>server

[Kermit Server running  on  the  DEC  Host.   Please  type  your  escape
sequence   to  return  to  your  local  machine.  Shut down the server by
typing the Kermit BYE command on your local machine.]


5) Return to Kermit on the RPi by typing <ctrl-\c> (that's control+backslash followed by the letter c). You should not be back at the RPi terminal window


----------------------------------------------------
(/opt/myfiles) C-Kermit>set file type text
(/opt/myfiles) C-Kermit>send myfile.txt

<<<There will be a whole page of file sending status>>>


6) If you want to recieve the .txt file instead of sending it, use

(/opt/myfiles) C-Kermit>get myfile.txt


7) To gracefully close down the connection use <ctrl-\c> (that's control+backslash followed by the letter c) to go back to Tops-10.

(/opt/myfiles) C-Kermit><ctrl-\c>
 DNS Lookup...  Trying 127.0.0.1...  Reverse DNS Lookup... (OK)
Connecting to host localhost:2020
 Escape character: Ctrl-\ (ASCII 28, FS): enabled
Type the escape character followed by C to get back,
or followed by ? to see other options.
----------------------------------------------------


Connected to the PDP-10 simulator DZ device, line 1

<<<Log out from Tops-10 here>>>

<ctrl-\c>

----------------------------------------------------
(/opt/myfiles) C-Kermit><ctrl/c>

pi@raspberrypi:/opt/myfiles $


-- steve

[Ric Werme]

On TOPS-10, your kermit is version 3(136).  I think I found [almost] that at Columbia a week or two ago.  What I have on the stock PdDP-10 distribution is 5A(44)-7.  It seems to be a fairly horrible piece of code.  In experimenting with various commands gleaned from the V3 doc,  several don't work.

.dir kermit.*[1,*]

KERMIT  EXE   136  <155>   21-Apr-78    5A(44)-7        DSKB:   [1,4]

.r kermit

[Kermit Ver. 5A(44)-7]

Kermit-10> set file byte-size seven-bit

?KERSET (FILE) Illegal arguement [sic]
?KERSET (SET) Illegal or ambigious [sic] keyword

?SCNEXA Excess arguments starting with " " following word "SIZE"

Kermit-10> dir
HELLO   EXE?
?HALT at user PC 000000

.dir

HELLO   EXE    14  <057>   12-Sep-78    DSKB:   [10,407]
README  TXT     0  <057>   20-Aug-78
LIGHT   REL     3  <057>    4-Sep-78
...
.

I downloaded Columbia's code but haven't looked at it closely other than seeing it sure isn't what I'm running.  Here's a bit of what may be a bit older than what you're running:

ss:k10$ ls -l | head
total 1872
-rw-r--r-- 1 werme werme   2580 Apr 26  2006 k10133.mem
-rw-r--r-- 1 werme werme   2343 Apr 26  2006 k10133.rno
-rw-r--r-- 1 werme werme   3768 Apr 27  2006 k10.ann
-rw-r--r-- 1 werme werme    182 Apr 26  2006 k10bli.ccl
-rw-r--r-- 1 werme werme   6191 Apr 26  2006 k10com.req
-rw-r--r-- 1 werme werme    588 Apr 26  2006 k10err.r36
-rw-r--r-- 1 werme werme   4510 Apr 26  2006 k10glb.bli
-rw-r--r-- 1 werme werme   8608 Apr 26  2006 k10glb.mac
-rw-r--r-- 1 werme werme    132 Apr 26  2006 k10mit.bwr
ss:k10$ grep -a argument *.* | wc
    322    3100   21500
ss:k10$ grep -a arguement *.* | wc
      0       0       0
ss:k10$ grep -a ambig *.*
ss:k10$ grep -a PATCH k10133.mem
PATCH [127]
PATCH [130] & [3.1.045]
PATCH [131], [132], & [133]
PATCH [134]

So, the obvious question is - Where did you get your Kermit?  Is with your 7.03 distribution?

Alternatively, WTH is the kermit we get with the PiPD-10's release?

I guess I can try compiling this code, it may use some KI10/KL10-only instructions, that would be easy to fix, I don't think I need a BLISS compiler.

I'm falling behind on other stuff I need to do, so haven't been spending too much time on this lately.

[steve...@gmail.com]

"So, the obvious question is - Where did you get your Kermit?  Is with your 7.03 distribution?"

The original source of the 7.03 image that I use probably came from "Quentin's Blog" and had that version of Kermit installed:

https://www.quentin.org.uk/2018/02/15/configuring-pdp10-tops-10-for-online-use/

He talks about " pre-configured Tops-10 disk packs", that's what I started with, I believe.

The only other place my Tops-10 7.03 disks could have come from would have been from the Retrocmp project's PanelSim simulation of a KL-10 although I think they also got their disk images from Quentin's Blog. I got lots of practice on Tops-10 and the PDP-10 playing with Retrocmp's KL-10 simulator while waiting for Oscar's PiDP-10 to become available. Once the PiDP-10 was available I just moved my disk images over, finding that Tops-10 7.03 doesn't run on a KA-10 CPU so I had to tweak things to get the PiDP-10 to boot into a KS-10 CPU.

-- steve

[Noah Smith]

hi all, the following could be pretty far off the path, so use with caution:) - just throwing it out there fwiw.

for decwar ended up scripting creation of kl-10 disk from scratch, including mongen etc. long story. bypasses the common disk images floating around.

got that going by building on rich's fantastic work - rich made all of that possible.

as part of that, kermit does get installed from tape and ends up working fine.

don't remember all the details, but here are notes - and the resulting overall decwar environment / disk does have a working kermit
https://gitlab.com/decwar/utexas/-/blob/main/docs/sec6-scripted-tops10-setup.md

cheers,

noah

[terri-...@glaver.org]

On Saturday, October 11, 2025 at 3:57:45 PM UTC-4 r...@wermenh.com wrote:

On TOPS-10, your kermit is version 3(136).  I think I found [almost] that at Columbia a week or two ago.  What I have on the stock PdDP-10 distribution is 5A(44)-7.  It seems to be a fairly horrible piece of code.  In experimenting with various commands gleaned from the V3 doc,  several don't work.

Columbia University shut down its Kermit efforts quite a few years ago. The Kermit Project is now independent and can be found at https://www.kermitproject.org 

The latest PDP-10 Kermit is 3(136) from April 2006. There's also a Lisp implementation for ITS, developed elsewhere. For completeness, there's a DECSYSTEM-20 Kermit 5.3(277)-5, 30 August 2024.

Searching the entire Kermit Project web site, "5A(44)-7" does not appear at all. A general google search for "5A(44)-7" returns no relevant matches, and none at all if you add "kermit" to the search string. So I have no idea where that came from.

[Ric Werme]

Oh - The code I have came from the Kermit Project, https://www.kermitproject.org/pdp10.html .  3(136) may be a typo, I see other references to 3(134), both in the download code and at https://www.kermitproject.org/archive.html OTOH, the latter refers to date 1989/09/08 and the files in the tar ball have an Apr 26  2006 date.

In https://www.kermitproject.org/ck90.html the "Individual binaries" lists these oldest entries:

188: Version 5A(188), November 1992 through September 1993.
189: Version 5A(189), September 1993 through October 1994.
190: Version 5A(190), October 1994 through September 1996.
192: Version 6.0.192, September 1996 through December 1999.

So count that as a hint about 5A(44)-7.

[terri-...@glaver.org]

On Saturday, October 11, 2025 at 9:42:54 PM UTC-4 r...@wermenh.com wrote:

Oh - The code I have came from the Kermit Project, https://www.kermitproject.org/pdp10.html .  3(136) may be a typo, I see other references to 3(134), both in the download code and at https://www.kermitproject.org/archive.html OTOH, the latter refers to date 1989/09/08 and the files in the tar ball have an Apr 26  2006 date.

The https://www.columbia.edu/kermit/ftp/d/k10.ann file definitely calls it 3(136). That page is linked from the https://www.kermitproject.org/pdp10.html page, so is likely current.

 

In https://www.kermitproject.org/ck90.html the "Individual binaries" lists these oldest entries:

188: Version 5A(188), November 1992 through September 1993.
189: Version 5A(189), September 1993 through October 1994.
190: Version 5A(190), October 1994 through September 1996.
192: Version 6.0.192, September 1996 through December 1999.

So count that as a hint about 5A(44)-7.

  

Could be. But a version 5 of C-Kermit would be ancient and it doesn't seem to (at least currently) support the 36-bit DEC systems. It would be interesting to see the output of "help", "show version" or similar when running the 5A(44)-7 binary. That might provide info on who / how / when / where it was built.

[Ric Werme]

> It would be interesting to see the output of "help", "show version" or similar when running the 5A(44)-7 binary.

Yes, it would be.  Unfortunately:

.r kermit

[Kermit Ver. 5A(44)-7]

Kermit-10> help   [This might merely  print a file like kermit.hlp, but I don't see anything like it.]
Kermit-10> show all
?KERCMD (SHOW) Illegal or ambigious keyword
Kermit-10> show
?KERCMD SHOW What?
Kermit-10> show file
?KERCMD No attribute specified to show
Kermit-10> show file all
?KERCMD Illegal or amgibious [sic - one of at least two missspellings] keyword

Kermit-10> version
[Kermit Ver. 5A(44)-7]
Kermit-10> usage
DSK:    Used:?

?HALT at user PC 000000

.r kermit

Kermit-10> daytim
12-Oct-78  0:34:34
Kermit-10> ddt
?KERCMD No DDT loaded
Kermit-10> show help
?KERCMD (SHOW) Illegal or ambigious keyword
Kermit-10> show debug
DEBUGGER:             OFF
Kermit-10> show eightb
EIGHT-BIT Mode:       OFF       EIGHT-BIT Char:         &

Some of these commands were gleaned from my new "fildmp" program.  It's useful!I I may write a "more" (or "less) or page command to step through the octal/sixbit/ascii dump.

[Ric Werme]

Well, I finally had time to make a start at building Kermit on 6.03.  I didn't get very far:

.copy k10unv.mac=ptr:

.comp k10unv.mac
MACRO:  KERUNV

?MCRCFU CANNOT FIND UNIVERSAL GLXMAC

EXIT

.

It appears that what's needed from 7.01 isn't on 6.03.  From k10mit.ctl:

! The following are required from various Digital distribution tapes
!
! Software built from sources on the 7.01A CUSP tape
!
!       GLXMAC.UNV - Universal GLXLIB (Galaxy library)
!       ORNMAC.UNV - Universal for OPR command interface
!       GLXLIB.REL - Loadable GLXLIB library
!       OPRPAR.REL - Loadable OPR command module

I'm going to put all this on the shelf.  Perhaps I'll play around sometime with mounting the 7.01 distribution with the 6.03 PiDP-10 distribution, but it's time to turn my attention to ITS and the 340 display.

Well,  after some rabble rousing about the memory display on the PiDP-10.  That really needs to get out to the community.