You mean have something external sitting on the CTY, typing into it and
not bothering to determine where all this stuff goes in Tops-10?? Well,
that's certainly one way to skin that particular cat...
But I don't remember having to do much of anything to Tops-10 to get it
to come up except maybe type the date and time in case the PDP-11 didn't
happen to have it.? I forget about the KA and KI; I think they needed a
bootstrap toggled in, but this could be left in low core.
I don't think it would be a terrible idea to understand just a little
bit more to know what start up files to edit, but I certainly do
understand not wanting to be bothered with something.? If I remember any
more, I'll let you know.
The new SIMH port appears to allow for a slaved PDP-6, which I remember
seeing on the 9th floor (It was connected to the MIT AI KA-10).? I think
they only shared one moby.? The last I heard was that the KL sources to
ITS (MC) were lost.? Anyway, if the SIMH KL simulator allows multiple
CPU's, then you could run Tops-10 SMP, which really was a tour de
force.? Extremely cool.
------------------------------------------------------------------------
On 3/10/20 10:41 PM, Supratim Sanyal wrote:
31.37 (TWONKY) is just a straight TWONKY distribution on KLH-10. All
required keyboard interactions to get it to boot up are consistently
the same; so I might be able to wrap it up around an expect script ...
worth a shot.
------------------------------------------------------------------------
On 3/10/20 9:46 PM, Thomas DeBellis wrote:
I think you're right, but it has been _decades_ since I last used
Tops-10.? At WPI, we had a KA-10 running a much modified 6.03 series
monitor that we were quite proud of.? At Marlboro, the project that I
was working on (FILE-FINDER, a database for DUMPER tapes) was quite
Tops-20 centric; we depended on files with holes in them.
I'm unaware of any systems level structured data store in either
Tops-10 or Tops-20 with the exception of the Quasar failsoft file
(QSRFSS, holds queue, print, batch requests across crashes).? I don't
find this surprising; if you crash and corrupt a file with
confuration information in it, a flat ASCII file is whaaay easier to
recover than an specially engineered database.? The binary accounting
and error files are sequential and don't count, IMHO.
Under Tops-20, we used the following 'trick' for start-up speed and
persisted configuration.? The configuration file was 'compiled' into
binary and directly mapped into memory on start-up.
1. This was necessary for LPTSPL as it is started up for jobs, but
shut down and put into a quiescent state when there is nothing
left to print.? When you have a lot of printers, reparsing
LPFORM.INI can be a real dog.? Very noticeable.
2. I got the idea from the mailer, which does the same thing for
mailing-list.txt
3. The EXEC will also do it; you can restore a binary environment
with all your special scripts really fast (like on PUSH or LOGIN)
4. I had been thinking about doing this for the Extended Mode FTP
server, but I'm not sure it's worth it.? I instrumented the start
up time and it's in the milliseconds.? Probably would be
necessary for a couple hundred simultaneous small requests.
If I ever get truly serious about supporting Galaxy again, then
probably I'll bite the bullet and put up Tops-10 so I can validate
execution.
> ------------------------------------------------------------------------
> On 3/10/20 9:25 PM, Johnny Billquist wrote:
>
> I've never used Tops-10 as an operator, so I can't answer most of
> this, but one question I think I can...
>
> My understanding is that neither Tops-10, nor TOPS-20 have a
> persistent database. Instead you need to have a script that does all
> the definitions, and you need to run it at every boot. But I could
> be confused about that one.
>> ------------------------------------------------------------------------
>> On 2020-03-11 01:50, Supratim Sanyal wrote:
>>
>> KLH-10 TOPS-10 noob questions:
>>
>> 1) At the TOPS-10 boot startup option prompt, I can type in CHANGE
>> and then set the DECnet address. How do I make it persist across
>> reboots and not have to do this every time?
--
Supratim Sanyal, W1XMT
39.19151 N, 77.23432 W
QCOCAL::SANYAL via HECnet