On 10 Dec 2021, at 10:42, Marco Antoniotti wrote:

Got the MTS Volume 7 PDF.? Looks like all that is needed is in there.? There only the issue of "library organization" and file naming (especially w.r.t. the %INCLUDE statements).

Look at page 58 of Vol 7 where it talks about source libraries. I see that this description assumes you know the format of a source library in MTS. This may have been a reasonable assumption in 1982 when it was written, but probably not so much so 40 years later This library format was first used for the assembler (sometime around 1968) and is described starting around page 51 of Volume 14 which is available at . You can, of course, construct a library by hand using, e.g., the editor. There is also a program, *MacUtil, which will manipulate such a library. This is described starting on page 279 of Volume 14. I'm not sure how useful it would be for PL/I source libraries.

Mike

Thanks Mike,

Got the MTS Volume 7 PDF.? Looks like all that is needed is in there.? There only the issue of "library organization" and file naming (especially w.r.t. the %INCLUDE statements).

Holidays project :)

I will pester you guys here then.

All the best
Marco

开云体育

Hi

I just cooked up the code you can see at as an exercise in learning PL/I (using the (F) compiler, which is the one available in MTS, as far as I understand).
The question I have is how would you structure your code library in MTS and what would be the necessary incantations to compile and run the test program.

Any comment on the code and any help for MTS would be appreciated.

Meanwhile, here is a tantalizing, at least for some, screenshot.

Thank you

Marco

开云体育

I had no intention of having it be easy. If it's possible, id like to know how.

Yes, it's possible, but I wouldn't recommend trying it. I changed the host ID from MT to MA on my system and it was surprisingly difficult. I didn't try changing the banner from University of Michigan to anything else. That should be possible too, but I don't know what's involved. Back in 1988 new MTS sites were a rare event and changing the host ID and site name was far from the most difficult problem to be solved. Now it would make sense for it to be easier, but it's not high on the list.

Mike

Hi.
I'm running MTS, and noticed that the default hostname for the system distribution? from MTS Archive is (MT)
Is there a way I can change the hostname? The login screen also shows "University of Michigan Computing Center", is it possible to change this?
I can't get at the source files very well, and I would prefer an easier solution to customize these things.

On Mon, Apr 26, 2021 at 09:13:01PM -0700, marcoxa@... wrote:

Also, both SDL and Spinhawk already have support for TCP/IP.

May I suggest realizing that the TCP/IP support already in SDL and Spinhawk
is most likely completely unrelated to the special hardware box MTS used to
interface to the Internet. :-)



--

Kevin



Bruceville, TX

What's the definition of a legacy system? One that works!
Errare humanum est, ignoscere caninum.

On 27 Apr 2021, at 0:51, Fish Fish wrote:

Wholeheartedly agree! The last thing we need is yet another Hercules!

When you're done, Mike, simply publish your Hercules fork with your changes and together we'll make every effort to incorporate them into SDL Hyperion.

Great, that's exactly what I intend to do. Right now I'm trying to figure out how to get TCP and UDP servers to work. For example the FTP client works fine and you can connect from MTS to another FTP server, but I can't get the server started properly. Some file is missing or not configured right.

Mike

Marco wrote:

[...]

May I suggest NOT to release yet another Hercules version?
The extra bits could/should be folded into SDL *and* Spinhawk.

Wholeheartedly agree! The last thing we need is yet another Hercules!

When you're done, Mike, simply publish your Hercules fork with your changes and together we'll make every effort to incorporate them into SDL Hyperion.

--
"Fish" (David B. Trout)
Software Development Laboratories

mail: fish@...

Hi

thanks for the update...

May I suggest NOT to release yet another Hercules version??? The extra bits could/should be folded into SDL *and* Spinhawk.

Also, both SDL and Spinhawk already have support for TCP/IP.

Cheers

Marco

On 26 Apr 2021, at 6:22, Paco Linux wrote:

Are you planning to release this version of MTS and the modified Hercules? Is so, do yo have an estimated date?

Yes we'll release this version of MTS. We'll also release a Hercules that runs it one way or another, either as part of a regular version or separately. As for timing, it's a bit hard to say. It's been on my list for years, but I'm actually working on it now. How quickly it gets done depends on how many problems I find. Hopefully within a few months at least.

Mike

--

On 24 Apr 2021, at 3:48, marcoxa@... wrote:

is there any possibility of running a FTP server on MTS (SDL Hercules 4.2.x).? It works with TK4-

Not in the D6 version of MTS. That version of MTS is from 1988 and doesn't have much Internet support. The later, 1996, version does support FTP as well as other services and I've been working on getting this to work in Hercules. MTS used a special hardware box to interface to the Internet and a special version of Hercules that emulates this hardware will be required to make the 1996 MTS connect to the Internet. I've got much of that working and I'm sure the rest can be done without too much trouble.

Mike

Hi

is there any possibility of running a FTP server on MTS (SDL Hercules 4.2.x).? It works with TK4-

All the best

Marco

On 1 Feb 2021, at 10:36, Matteo Panella wrote:

recently I tried to play around with MTS again and I decided to set it up using Hercules 4.3. Long story short: after a few minutes the OPERATOR task on DS01 dies and asks for a 9-track tape for dumping (after cancelling the dump and restarting the task from CON1 it seems to work without too much trouble).

I finally ran into this myself. I took a look at the jobdump produced (it's the first time I've done that in over two decades) and I think I know what is happening. In some rare cases Hercules is returning zero bytes of data for a 3270 "read modified" command. This is not legal. A read modified should always return at least 3 bytes (an AID and two bytes of cursor location). The operator job crashes if a read modified returns zero bytes. So there are really two bugs here: Hercules shouldn't return zero bytes for read modified and the operator job shouldn't crash if it does. The fix for the operator job is simple, but I don't think it is easy (or even possible) for you to rebuild it. I'll see if I can find why Hercules is doing this and see if there is a simple fix.

Mike

On 1 Feb 2021, at 10:36, Matteo Panella wrote:

recently I tried to play around with MTS again and I decided to set it up using Hercules 4.3. Long story short: after a few minutes the OPERATOR task on DS01 dies and asks for a 9-track tape for dumping (after cancelling the dump and restarting the task from CON1 it seems to work without too much trouble).

This only happens on Hercules 4.2 and later, so I went ahead and began bisecting the Hercules code until I landed on the commit that seems to trigger the task crash:?

That commit makes a lot of changes to 3270 emulation so it seems likely that it is the cause of the problem. It turns out that I just started trying to get my MTS on Hercules setup back in shape so I could take care of some loose ends. I have work to do to add the MTS changes (mainly Him emulation) to SDL-Hyperion, but hopefully I'll get something working before too long. I'll keep my eyes open for this issue. Thanks for doing the work to find the commit that breaks things.

Mike

On Tue, Jan 19, 2021 at 09:04:47AM -0800, Drew Derbyshire wrote:

On 1/18/21 11:00 PM, Mark L. Gaubatz via groups.io wrote:

Interesting rates ? especially the commercial rates. At these rates, my
work on a single five-month project alone at the time would have had a
direct cost of roughly $7MM, and over $640MM in resources for QA (we were
running 16 engines, and sometimes more, 24x7, on machines significantly
faster than the 3090 at the time).

What kind of machine?

Were you the commercial rate?? Academic billing can be funny.

I know at Clarkson in late 1970's, there two type of charges on the
S/630-65J under OS/360: "funny" and "real" money.?? Most charges for routine
use were funny, but physical supplies like paper and cards were real.

A few years earlier ( early 1970's ), at the University of Alberta, our
360/67 running MTS, had a "system" called SOB Balls ( Student Oriented
Batch ), a funny money system, where students were alotted clear plastic
balls, smaller than ping-pong balls, which were handed in at the window
where they submitted their decks of cards and received their printed
output, one ball in exchange for one run.

More senior classes and instructors had terminal access, either through
Decwriters or glass terminals of various sorts ( AJ's being one ).
Those communicated through the PDP-11 front end machines.

Being a student at the time, I didn't have any contact with real
billing. As well as all of the academic work on that machine that
happened during the day time, the machine was switched to MVS during the
middle of the night to run administrative batch jobs.

The 360/67 was replaced with an Amdahl 470 and then 580, still running
MTS.



Brian McCullough