开云体育

ctrl + shift + ? for shortcuts
© 2025 开云体育

A number of questions about MTS.


 

Hello MTS community! After some moderate reading and light usage of
MTS, I would appreciate if you could answer some of my questions.



1.) Now that archive.michigan-terminal-system.org has been down for
almost a year - though well-archived on Archive.org, and with the
Google Drive links still existing - will this mailing list be the only
home of the MTS community, or will there eventually be another website?



2.) Obviously not everything can be released for legal/licensing
reasons, and sorting thru it all for D7.0 will take however long it
takes.

But I would like to know: are these files picked out yet but
just not publicly available pre-usable-system, or are most of these
files still being negotiated with rightsholders/programmers/etc? (As
far as I've read many/most of the post-D6.0 redistribution tapes are
still lost, so obviously those can't be sifted through yet.)



3.) Is it possible to connect a VT52/VT100/Tektronix to the current
D6.0A? Or would that require the networking support in D7.0? (Vol. 4
Appendix B reads to me like UMnet needs to be functioning first.)



4.) Browsed through CCM 448 [1] about VSS. However, I still have a few
questions:

a.) What version of MVS/370 did support stop at? Can it run
programs for the venerable 3.8j?

b.) Was there networking support?

[1]



5.) I read the previous message about the HIM merge. Is this fork
publicly available?



6.) There were several MTS systems on the BITNET network[1]. Looking
through the bitnet writeup (6.0T5 -1454), MTS BITNET supported
operation over multiple link protocols:

313 The full complement of RSCS link protocols will be
supported.
314 This requirement allows interworking with RSCS emulation
packages
315 which only implement a single protocol.


718 Link protocols (VMB, VMC, NJE, and X.25)


1166 At this writing, only the VMB DSP has been written at the
protocol
1167 layer and the SDA DSP at the physical device layer. In
addition,
1168 a subtasking monitor unit check routine for SDA device type
has
1169 been written.


1367 UBC and SFU have expressed a desire to implement their
Netnorth
1368 link using X.25 and Datapac, because bisync is not
available. It
1369 should be possible to substitute the UBCnet X.25 DSP for
either
1370 the SDA or CTCA DSPs to provide this function, in effect
1371 substituting a full duplex X.25 virtual circuit for a
permanent
1372 bisync line. The VMB or NJE protocols could then operate
over
1373 this link.

Hyperion uses a 2703-lookalike[2] to tunnel NJE over TCP/IP. MTS also
supported the 2703[3], although it was eventually replaced with the
PDP-8 Data Concentrator[4], which was itself replaced with the PDP-11...

-----> The point: did the PDP-11 HIM/NIM/PCP/SCP/whatever support
NJE over a bisync line? Or was it TCP/IP, X.25, and local terminals
only? Would the 2703 be the only way to get BITNET on MTS, or
would an NJE-supporting frontend be the better option, perhaps in
combination with NJE-Bridge[5]? <-----

[1]
[2]
[3]
[4]
[5]



7.) ASMH is not available due to licensing. More than half the
system is written in Plus[1]. Is it possible to recompile only the
Plus-based portions, plugging in the already-assembled (and/or
manually patched) ASMH portions after the fact? Or is ASMH still
required for the existing build process, absent some heavy
modifications to the build scripts? (Not that I'm capable of any
meaningful modifications, as my programming skills are nascent.)

[1]



Thank you for for taking time to respond to these questions.


 

开云体育

I'll try to answer some of these inline below. Let me know what else you need.

On 22 Jan 2024, at 1:10, Bile Geek wrote:

Hello MTS community! After some moderate reading and light usage of
MTS, I would appreciate if you could answer some of my questions.

1.) Now that archive.michigan-terminal-system.org has been down for
almost a year - though well-archived on Archive.org, and with the
Google Drive links still existing - will this mailing list be the only
home of the MTS community, or will there eventually be another website?

Yes, there will be a new web site. We were just talking about this today.

2.) Obviously not everything can be released for legal/licensing
reasons, and sorting thru it all for D7.0 will take however long it
takes.

But I would like to know: are these files picked out yet but
just not publicly available pre-usable-system, or are most of these
files still being negotiated with rightsholders/programmers/etc? (As
far as I've read many/most of the post-D6.0 redistribution tapes are
still lost, so obviously those can't be sifted through yet.)

All of the so-called "redistribution" tapes were lost. These were informal tapes shared among the MTS sites (sort of like Github without the internet). The rest of the stuff is still being sorted through.

3.) Is it possible to connect a VT52/VT100/Tektronix to the current
D6.0A? Or would that require the networking support in D7.0? (Vol. 4
Appendix B reads to me like UMnet needs to be functioning first.)

This is a complicated question and the answer depends on exactly what you want to do. Connecting a real glass teletype to MTS (i.e. a VT100 or something that emulates one) will be very difficult. However if you run MTS in Hercules you can connect a tn3270 client to MTS right now. Hercules turns a tn3270 connection into a local 3270 which MTS supports. You probably could hack together something that would connect a VT100 emulator to MTS, but I don't think anyone has tried this and I'm not sure how you would do it.

4.) Browsed through CCM 448 [1] about VSS. However, I still have a few
questions:

a.) What version of MVS/370 did support stop at? Can it run
programs for the venerable 3.8j?

Something old. I don't think any particular version was targeted and the emulation was not complete. Mainly it emualted anything needed by a program we wanted to run. It was never meant to be a complete emulation of MVS.

b.) Was there networking support?

No

[1]

5.) I read the previous message about the HIM merge. Is this fork
publicly available?

No, not yet. Soon, I hope.

6.) There were several MTS systems on the BITNET network[1]. Looking
through the bitnet writeup (6.0T5 -1454), MTS BITNET supported
operation over multiple link protocols:

313 The full complement of RSCS link protocols will be
supported.
314 This requirement allows interworking with RSCS emulation
packages
315 which only implement a single protocol.

718 Link protocols (VMB, VMC, NJE, and X.25)

1166 At this writing, only the VMB DSP has been written at the
protocol
1167 layer and the SDA DSP at the physical device layer. In
addition,
1168 a subtasking monitor unit check routine for SDA device type
has
1169 been written.

1367 UBC and SFU have expressed a desire to implement their
Netnorth
1368 link using X.25 and Datapac, because bisync is not
available. It
1369 should be possible to substitute the UBCnet X.25 DSP for
either
1370 the SDA or CTCA DSPs to provide this function, in effect
1371 substituting a full duplex X.25 virtual circuit for a
permanent
1372 bisync line. The VMB or NJE protocols could then operate
over
1373 this link.

You will note that much of this is written in the future tense. At the time of D6 the Resource Manager (RM) was not yet integrated into MTS so none of this is available in D6. The 1996 system has the RM and it works well. I haven't tried BITNET since there is no one for me to talk to and configuring it is a pain. I have however gotten a Postscript printer on my local net to print from MTS using the RM.

Hyperion uses a 2703-lookalike[2] to tunnel NJE over TCP/IP. MTS also
supported the 2703[3], although it was eventually replaced with the
PDP-8 Data Concentrator[4], which was itself replaced with the PDP-11...

-----> The point: did the PDP-11 HIM/NIM/PCP/SCP/whatever support
NJE over a bisync line? Or was it TCP/IP, X.25, and local terminals
only? Would the 2703 be the only way to get BITNET on MTS, or
would an NJE-supporting frontend be the better option, perhaps in
combination with NJE-Bridge[5]? <-----

Your description of the history of terminal support in MTS is mostly correct except that the 2703 wasn't replaced by any of the others, it coexisted with them until the end. I'm not an expert on BITNET by a long shot and BITNET was never more than a bump on the side of MTS. I know we supported EMail via BITNET but I can't recall what other services, if any, were supported over BITNET. What do you want to use BITNET for? Are there any hosts out there still using it?

[1]
[2]
[3]
[4]
[5]

7.) ASMH is not available due to licensing. More than half the
system is written in Plus[1]. Is it possible to recompile only the
Plus-based portions, plugging in the already-assembled (and/or
manually patched) ASMH portions after the fact? Or is ASMH still
required for the existing build process, absent some heavy
modifications to the build scripts? (Not that I'm capable of any
meaningful modifications, as my programming skills are nascent.)

You should be able to rebuild any module (Plus, ASMH, GOM, or whatever) without affecting any other module. Neither the GOM nor the Plus compiler depends on ASMH.

[1]

Thank you for for taking time to respond to these questions.


 

You probably could hack together something that would connect a VT100
emulator to MTS, but I don't think anyone has tried this and I'm not
sure how you would do it.
That's exactly what I wanted to do. Oh well; thanks for clarifying!

(...not that it would be the best experience (the UI is definitely
3270-centric, been using x3270 myself), but I just wanted to try it out
for completeness' sake.)



You will note that much of this is written in the future tense. At
the time of D6 the Resource Manager (RM) was not yet integrated into
MTS so none of this is available in D6. The 1996 system has the RM
and it works well. I haven't tried BITNET since there is no one for
me to talk to and configuring it is a pain. I have however gotten a
Postscript printer on my local net to print from MTS using the RM.
Speaking of printers: do you recall if later versions of LaTeX[1]
supported PostScript, or if support was stuck on the Xerox 9700?

[1]



Your description of the history of terminal support in MTS is mostly
correct except that the 2703 wasn't replaced by any of the others, it
coexisted with them until the end.
Ah, that's good news! Less hassle to setup then.



What do you want to use BITNET for? Are there any hosts out there
still using it?
Yes, there's a hobbyist BITNET network called HNET[1][2]. Though making
a local-only network would be cool too.

Would also be interesting to connect the HIM to the TELEBAHN, but
obviously X.25 support has far less priority than TCP/IP.

[1]
[2]
[3]/g/x25


 

Dangit, I always come up with one more question after hitting send...

Memory is obviously cheap these days, so I could give Hercules 9gb (and
have!) without breaking a sweat. However, I'm curious about how small I
could go on MAINSIZE (on either D6.0A or the future D7.0) and still
have a usable system.

Obviously memory requirements would be bigger for a system that has
many users doing actual work, but would say 32mb be sufficient for a
handful of hobbyist users? How much memory would be required for
development work on the system itself, editing and recompiling files?


 

I promise this is the last time spamming your / the group's inbox, just
qualifying to the above email:

*how low can MAINSIZE go [without hitting swap constantly, or
even often]

People have obviously run timesharing systems on kilobytes of RAM, just
wondering what's comfortable for later MTS.


 

We never had anywhere near 32 MB on an MTS machine. I think the biggest one might have been 8 MB and we supported up to 300 users on that machine (although it was sometimes a struggle). Whether you would get paging on a 32 MB machine depends, of course, on what people are doing.
I can bog down the paging system all by myself if I want to. However a few people doing "normal" things won't put a significant paging load on the machine.

Editing files puts almost no load on the machine. Compiling them is more intensive, but still not very. I don't recall that ever being an issue. If the machine was slow it wasn't due to someone compiling something.

Mike

On 26 Jan 2024, at 21:02, Bile Geek wrote:

I promise this is the last time spamming your / the group's inbox, just qualifying to the above email:

*how low can MAINSIZE go [without hitting swap constantly, or even often]

People have obviously run timesharing systems on kilobytes of RAM, just wondering what's comfortable for later MTS.


 

I have a (very dim) recollection that the 3083 at SFU in 1991 had a whopping 24 MB of memory - but someone else from SFU can correct me if I'm hallucinating again. ;-)

It's fun to run MTS on a modern machine and give it 128 MB of memory (because I can!) and never have to page. Even when I had it on a Raspberry Pi 1. :-)

For serial devices - if someone can generate a system with a 2703, async (and bisync?), would it be possible to get 'physical' async and bisync to work? Or to emulate? I used to have an ITEL/Dura Selectric terminal which would have been fun to watch on MTS again, but even it got consigned to the scrap heap years ago. I do have a couple of old PCs with serial terminal emulators on them, including a multi-window VT100 emulator that I wrote in Turbo Pascal when working on the NIMs.

For bisync, 2780 protocol is trivial, and would be easy to emulate - if it isn't already out there in the form that you want. It would probably take a day to make work. 3780 is a little more complex, but still easy.

2703 provided very basic I/O even to video terminals. You could get graphics on a Tektronix 40xx screen, but there was no full screen editing or such. For the UBCnet-style networking you would need a Node and a NIM (hardware or emulated) to get async full-screen terminal support - but x3270 exists now and works. I had thought of running a Node and a NIM in a PDP11 emulator as a masochistical exercise - just to see if it works- but there would still be work in emulating something like an Auscom channel interface in the PDP11 emulator as well, and a way to interface it to Hercules.

I have been using some crude scripts'n'tricks to get printouts and file transfers on D6.0 using the bare reader/printer/punch devices since MTS was released for Hercules - until 'real' networking is eventually available.

- Richard, VE7CVS

On 1/26/24 9:53 PM, Mike Alexander wrote:
We never had anywhere near 32 MB on an MTS machine.? I think the biggest one might have been 8 MB and we supported up to 300 users on that machine (although it was sometimes a struggle).? Whether you would get paging on a 32 MB machine depends, of course, on what people are doing.? I can bog down the paging system all by myself if I want to.? However a few people doing "normal" things won't put a significant paging load on the machine.

Editing files puts almost no load on the machine.? Compiling them is more intensive, but still not very.? I don't recall that ever being an issue.? If the machine was slow it wasn't due to someone compiling something.

Mike

On 26 Jan 2024, at 21:02, Bile Geek wrote:

I promise this is the last time spamming your / the group's inbox, just
qualifying to the above email:

*how low can MAINSIZE go [without hitting swap constantly, or
even often]

People have obviously run timesharing systems on kilobytes of RAM, just
wondering what's comfortable for later MTS.


 

On 26 Jan 2024, at 17:35, Bile Geek wrote:

Speaking of printers: do you recall if later versions of LaTeX[1] supported PostScript, or if support was stuck on the Xerox 9700?

I don't think I answered this question.

When MTS was shut down in 1996 Postscript support was mostly complete but not really released yet. However it mostly works.

MTS documentation is generally Textform input, LaTex input, or printer ready files meant to be copied to a printer.

There is a new version of Textform in NEW:TEXTFORM that will produce Postscript output.

DVI files produced by LaTex in MTS can be processed by dvips on a Mac (and presumably elsewhere). It's also possible to copy the LaText file to a Mac and run LaTex on it there. Of course the version of LaTex and TeX in MTS is 30 years old and somewhat incompatible with current versions.

I have a version of enscript that will turn an MTS printer file, complete with carriage control, into a Postscript file. Back around 2010 I used a combination of these techniques to convert much of the MTS documentation to PDF files. most of which are available online. There is a lot more that could be converted.

Mike


 

Mike,
One of the other questions was why are there no 2703 or similar lines
defined it the released versions.
I know when you did the release there was little support for these in
Hercules but we now have support for TELE2, 2741 and 3781 (bi-sync for
hasp).
These would be useful for TEK4010 or HPGL plotters, so could some be added?
Dave

-----Original Message-----
From: [email protected] <[email protected]> On Behalf Of Mike
Alexander
Sent: Wednesday, January 31, 2024 5:49 AM
To: [email protected]
Subject: Re: [H390-MTS] A number of questions about MTS.


On 26 Jan 2024, at 17:35, Bile Geek wrote:

Speaking of printers: do you recall if later versions of LaTeX[1]
supported PostScript, or if support was stuck on the Xerox 9700?
I don't think I answered this question.

When MTS was shut down in 1996 Postscript support was mostly complete but
not really released yet. However it mostly works.

MTS documentation is generally Textform input, LaTex input, or printer
ready
files meant to be copied to a printer.

There is a new version of Textform in NEW:TEXTFORM that will produce
Postscript output.

DVI files produced by LaTex in MTS can be processed by dvips on a Mac (and
presumably elsewhere). It's also possible to copy the LaText file to a
Mac and
run LaTex on it there. Of course the version of LaTex and TeX in MTS is
30 years
old and somewhat incompatible with current versions.

I have a version of enscript that will turn an MTS printer file, complete
with
carriage control, into a Postscript file. Back around
2010 I used a combination of these techniques to convert much of the MTS
documentation to PDF files. most of which are available online. There is
a lot
more that could be converted.

Mike




 

On 31 Jan 2024, at 12:04, Dave Wade wrote:

One of the other questions was why are there no 2703 or similar lines defined it the released versions. I know when you did the release there was little support for these in Hercules but we now have support for TELE2, 2741 and 3781 (bi-sync for hasp). These would be useful for TEK4010 or HPGL plotters, so could some be added?

Sure, some could be added. However we aren't really anxious to do more with D6. I'd rather put my effort into getting the 1996 system out.

Mike


 

-----Original Message-----
From: [email protected] <[email protected]> On Behalf Of Mike
Alexander
Sent: Thursday, February 1, 2024 7:37 AM
To: [email protected]
Subject: Re: [H390-MTS] A number of questions about MTS.

On 31 Jan 2024, at 12:04, Dave Wade wrote:

One of the other questions was why are there no 2703 or similar lines
defined it the released versions.
I know when you did the release there was little support for these in
Hercules but we now have support for TELE2, 2741 and 3781 (bi-sync for
hasp).
These would be useful for TEK4010 or HPGL plotters, so could some be
added?
Sure, some could be added. However we aren't really anxious to do more
with
D6. I'd rather put my effort into getting the 1996 system out.
Well in that case, could we have them in the 1996 system ?


Mike
Dave