[its-hackers] Fwd: [PDP-10/its] Use the CH11 Chaosnet address (#2038)

Jack Haverty jack at 3kitty.org
Wed Jun 23 20:26:04 CEST 2021


Thanks, that all makes sense.   I do remember SEC working on the ML 
disks; it was very convenient to have another machine around.

A few observations for historical sake..

- ENG probably meant ENGineering, which related to "system" stuff, e.g., 
ITS itself and perhaps major applications like Midas, DDT, Teco, etc.   
Everything was in English back then.   ENG in a directory name was an 
indicator to mere mortals to keep their hands off that directory.

- When you see things with huge offsets (e.g., JRST MEMSIZ-400), it's 
likely to be one of two situations.  Either it was an instruction that 
was patched in using DDT, to jump to code somewhere else that did what 
needed to be done, usually to clean things up on the way out of some 
action.  Or it could also have been a jump into a module that had been 
removed.

E.g., after Dectapes become obsolete, at one point someone removed the 
tape-handling code, but didn't necessarily find all of the places 
elsewhere that had a JRST to some useful section of the tape code.  I 
remember that we had some bugs in the disk code after the tape code was 
removed, because it had been written to jump into the tape code where 
some functionality already existed.

MIT-DM code did all come from AI.   So it's likely that the DM system 
utilities like DSKDMP diverged from the AI version, at least for a 
while, to deal with the unique hardware components in DM (DSK, the E&S, 
etc.)   That's probably why there was no source code for a while.

Most of the ITS system utilities originally came from TMRC - the Tech 
Model Railroad Club, where some of the AI system hackers had built them 
for controlling the railroad.   I don't remember if there was any effort 
to archive their software.   But that might be another place to look for 
really old code.   In addition to migrating to MIT-AI, some of the TMRC 
people also took it to DEC.

/Jack


On 6/23/21 10:12 AM, Lars Brinkhoff wrote:
> Jack Haverty wrote:
>> Just curious --- has anyone found source code for DSKDMP?   Back in the
>> mid-70s, I recall an incident where someone asked for a copy of DSKDMP to
>> use on some new machine, and was surprised to learn that there wasn't any
>> source file.   Maybe there once was such a file, but over the years DSKDMP
>> had been extensively modified and new features added, all by simply writing
>> new code in DDT and saving the binary image.   So there was no source file
>> for the DSKDMP we used on MIT-DM, at least at that time.
> I could well believe this was true at some point.  Unfortunately the
> backup records are not in a good shape, so we don't have much evidence
> to go on.  Here are all DSKDMP source files I can see that were backed
> up on DM.  This is just information from the .TAPE database; we don't
> have the files themselves.
>
>   SAENG   DSKDMP  109    1975-01-15 12:12:12
>   SAENG   DSKDMP  132    1976-04-23 07:36:56
>   SAENG   DSKDMP  134    1976-04-24 23:46:12
>   SAENG   DSKDMP  144    1976-08-09 01:40:50
>   SYSTEM  DSKDMP  151    1977-04-02 21:00:57
>   SYSTEM  DSKDMP  155    1978-08-22 02:16:00
>   SYSTEM  DSKDMP  154    1978-03-16 04:05:16
>   SYSTEM  DSKDMP  156    1979-09-05 20:45:27
>   SYSTEM  DSKDMP  157    1979-10-18 06:38:41
>
> The first observation I make is that DM stored source files in a
> directory called SAENG.  This is unique to DM.  I'm guessing it means
> "Stand-Alone English files".
>
> Secondly, there is a decent progression of file versions from 132 to
> 157.  There *is* a decided jump from 109 to 132 and timestamps have them
> separated by more than a year.  We can imagine this to be consistent
> with your mid-70s incident, but that's a rather flimsy assumption.
>
> I understand Scott Cutler was one of the people who worked on attaching
> disks to DM, using the new Mathlab PDP-10 for testing and bringup since
> it conveniently had the same kind of DEC disk controller.
>
> Outside Dynamic Modeling we have good coverage of source files for
> DSKDMP going all the way back to 1971 were it seems George Mitchell and
> Fred Wright were involved with (creating?) early versions.  It seems not
> long before that MACDMP was used to boot ITS.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://its.victor.se/pipermail/its-hackers_its.victor.se/attachments/20210623/bf620c5c/attachment.html>


More information about the its-hackers mailing list