Date   

Re: micro-emacs for MVS

Joe Monk
 

"A minimal control unit."

That's not a valid answer. You will need a control unit to translate CCW to whatever protocol this "EBCDIC ANSI" terminal will use... RS232? 3270?

Joe

On Wed, Dec 15, 2021 at 6:39 PM Paul Edwards <mutazilah@...> wrote:
On Thu, Dec 16, 2021 at 11:34 AM, Joe Monk wrote:

> "When my real EBCDIC ANSI terminal is bolted onto the
> fiber or bus/tag connection"
>
> Ok, how? Your "real" EBCDIC ANSI "terminal" will have to be connected to a
> controller, which will then talk to the channel.
>
> What control unit will you use?

A minimal control unit.

BFN. Paul.






Re: micro-emacs for MVS

Paul Edwards
 

On Thu, Dec 16, 2021 at 11:34 AM, Joe Monk wrote:

"When my real EBCDIC ANSI terminal is bolted onto the
fiber or bus/tag connection"

Ok, how? Your "real" EBCDIC ANSI "terminal" will have to be connected to a
controller, which will then talk to the channel.

What control unit will you use?
A minimal control unit.

BFN. Paul.


Re: micro-emacs for MVS

Joe Monk
 

"When my real EBCDIC ANSI terminal is bolted onto the
fiber or bus/tag connection"

Ok, how? Your "real" EBCDIC ANSI "terminal" will have to be connected to a controller, which will then talk to the channel.

What control unit will you use?

Joe

On Wed, Dec 15, 2021 at 4:21 PM Paul Edwards <mutazilah@...> wrote:
On Thu, Dec 16, 2021 at 09:06 AM, Dave Wade wrote:

> > What is the correct technical solution for what I am trying to do?
>
> I would use TCPIP.

When my real EBCDIC ANSI terminal is bolted onto the
fiber or bus/tag connection, there is no TCP/IP involved,
and no reason to add huge amounts of code to z/PDOS
to support it.

Within those constraints, what is the correct technical
solution?

Thanks. Paul.






Re: micro-emacs for MVS

Paul Edwards
 

On Thu, Dec 16, 2021 at 09:48 AM, Dave Wade wrote:

Why would anyone design and build a character mode terminal?
It's not really character mode, is it? Sometimes the
block will consist of a single character. Sometimes
it will be multiple characters. And that's just for input.
For output it could be a large block.

If you must I don't believe there is a "correct."
I suppose my only observation, is that, if my experience is to be gone by, you
will regret whichever you choose....
I don't think I've regretted any decision I have made so far
on MVS. Why would this be the first? As I see it, both
BSAM tape and 1052 EXCP are appealing, and whichever
one I go with, I won't have any knowledge to know that it
is wrong in order to regret it.

And everything is being peer reviewed too.

BFN. Paul.


Re: micro-emacs for MVS

Dave Wade
 

-----Original Message-----
From: hercules-380@groups.io <hercules-380@groups.io> On Behalf Of Paul
Edwards
Sent: 15 December 2021 22:22
To: hercules-380@groups.io
Subject: Re: [hercules-380] micro-emacs for MVS

On Thu, Dec 16, 2021 at 09:06 AM, Dave Wade wrote:

What is the correct technical solution for what I am trying to do?
I would use TCPIP.
When my real EBCDIC ANSI terminal is bolted onto the fiber or bus/tag
connection, there is no TCP/IP involved, and no reason to add huge amounts
of code to z/PDOS to support it.

Within those constraints, what is the correct technical solution?
With my IBM hat on my solution would be a 3270.Thats what IBM designed to go with its Mainframes.
Why would anyone design and build a character mode terminal?
If you must I don't believe there is a "correct."
I suppose my only observation, is that, if my experience is to be gone by, you will regret whichever you choose....
Sorry..

Thanks. Paul.
Dave


Re: micro-emacs for MVS

Paul Edwards
 

On Thu, Dec 16, 2021 at 09:06 AM, Dave Wade wrote:

What is the correct technical solution for what I am trying to do?
I would use TCPIP.
When my real EBCDIC ANSI terminal is bolted onto the
fiber or bus/tag connection, there is no TCP/IP involved,
and no reason to add huge amounts of code to z/PDOS
to support it.

Within those constraints, what is the correct technical
solution?

Thanks. Paul.


Re: micro-emacs for MVS

Dave Wade
 

Tony,

 

Thanks for that Info. I suspected that was the case. I did open the BTAM manual and quickly closed it

 

Dave

 

From: hercules-380@groups.io <hercules-380@groups.io> On Behalf Of Tony Harminc
Sent: 15 December 2021 21:15
To: hercules-380@groups.io
Subject: Re: [hercules-380] micro-emacs for MVS

 

On Wed, 15 Dec 2021 at 05:48, Dave Wade <dave.g4ugm@...> wrote:


No I think you would have to use EXCP. Doesn't BSAM in MVS 3.8J support any terminal devices?

 

No, it doesn't. (Unless you consider the printer/card-reader combo a terminal - such a pair can be used by MVS as a console).

 

The only access method that supports terminals is BTAM. (No, for these purposes QTAM and VTAM are not access methods.)

 

Tony H.


Re: micro-emacs for MVS

Dave Wade
 

-----Original Message-----
From: hercules-380@groups.io <hercules-380@groups.io> On Behalf Of Paul
Edwards
Sent: 15 December 2021 21:44
To: hercules-380@groups.io
Subject: Re: [hercules-380] micro-emacs for MVS

On Wed, Dec 15, 2021 at 11:25 PM, Dave Wade wrote:

There is no problem with my hacked 1052 except for the fact that
under z/PDOS I basically have BSAM support for a 1052, and that BSAM
support does not exist on MVS, so Hecules/380 allows an EBCDIC ANSI
terminal to run on z/PDOS but not on MVS, and I'd like a solution
that works on both environments.
So you must have some channel programs that work with this. Why don't
you tweak these to work with EXCP and wrap it up in a BSAM like layer
Excellent point. We haven't discussed the alternative to tape.

I have the BSAM-like layer already - mvssupa - so it's a matter of adding EXCP
to that. So that's a matter of finding someone with the skills to write 370
assembler. Then I would need to update PDOS to handle the EXCP SVC 0.
This may well be the correct technical solution - handle raw, unknown
devices. In actual fact, ironically, mvssupa already seems to do EXCP - for
tapes! Maybe it is a matter of modifying that code to make it do it for all
devices, not just tapes.
NO! At least not for DASD.


The next issue is at the 1052 terminal side. The fundamental problem
appears to be that the 1052 has an implied "CR"
I think thats the way the hardware works.

whenever you do a write, which is the thing I probably don't want. I think I
want a device that does pure passthrough, and if you want a CR, send it
yourself. PDOS has logic in it that detects if there is a NL in the string, and if
there is one, it issues the write CCW, but if there is no NL, it issues a "write
with no CR" CCW. I probably want to remove this logic.

Or alternatively, maybe I can detect a terminal and always issue a write with
no CR CCW.

What is the correct technical solution for what I am trying to do?
I would use TCPIP.

Thanks. Paul.
No probs. The department I worked for in Salford was called Computer Services Research and Development.
Our job was to come up with hacks so we could deliver services as opposed to pure research.

These days I seem to have lost my "of course It can be made it work" attitude ......

Dave
G4UGM


Re: micro-emacs for MVS

Paul Edwards
 

On Wed, Dec 15, 2021 at 11:25 PM, Dave Wade wrote:

There is no problem with my hacked 1052 except for the fact that under
z/PDOS I basically have BSAM support for a 1052, and that BSAM support
does not exist on MVS, so Hecules/380 allows an EBCDIC ANSI terminal to run
on z/PDOS but not on MVS, and I'd like a solution that works on both
environments.
So you must have some channel programs that work with this. Why don't you
tweak these to work with EXCP and wrap it up in a BSAM like layer
Excellent point. We haven't discussed the alternative to tape.

I have the BSAM-like layer already - mvssupa - so it's a matter
of adding EXCP to that. So that's a matter of finding someone
with the skills to write 370 assembler. Then I would need to update
PDOS to handle the EXCP SVC 0. This may well be the correct
technical solution - handle raw, unknown devices. In actual fact,
ironically, mvssupa already seems to do EXCP - for tapes! Maybe
it is a matter of modifying that code to make it do it for all
devices, not just tapes.

The next issue is at the 1052 terminal side. The fundamental
problem appears to be that the 1052 has an implied "CR"
whenever you do a write, which is the thing I probably don't
want. I think I want a device that does pure passthrough, and
if you want a CR, send it yourself. PDOS has logic in it that
detects if there is a NL in the string, and if there is one, it
issues the write CCW, but if there is no NL, it issues a "write
with no CR" CCW. I probably want to remove this logic.

Or alternatively, maybe I can detect a terminal and always
issue a write with no CR CCW.

What is the correct technical solution for what I am trying
to do?

Thanks. Paul.


Re: micro-emacs for MVS

Paul Edwards
 

On Thu, Dec 16, 2021 at 08:17 AM, Tony Harminc wrote:

In what sense does z/PDOS "support" BSAM? Are you saying a user program can
OPEN a BSAM DCB and then issue READ/WRITE/CHECK against it?
Yes, that is correct. That's all I actually need - for the code
in mvssupa.asm to work.

BFN. Paul.


Re: micro-emacs for MVS

Tony Harminc
 

On Tue, 14 Dec 2021 at 19:37, Paul Edwards <mutazilah@...> wrote:

> Yall are in la-la land...

Why do you say that? If I am running z/PDOS on z/Arch
hardware, there is no TCP/IP - I haven't written it. But
someone can manufacture an EBCDIC ANSI terminal.
z/PDOS supports BSAM so mvssupa.asm will continue
to work.

In what sense does z/PDOS "support" BSAM? Are you saying a user program can OPEN a BSAM DCB and then issue READ/WRITE/CHECK against it?
Tony H.


Re: micro-emacs for MVS

Tony Harminc
 

On Wed, 15 Dec 2021 at 05:48, Dave Wade <dave.g4ugm@...> wrote:

No I think you would have to use EXCP. Doesn't BSAM in MVS 3.8J support any terminal devices?

No, it doesn't. (Unless you consider the printer/card-reader combo a terminal - such a pair can be used by MVS as a console).

The only access method that supports terminals is BTAM. (No, for these purposes QTAM and VTAM are not access methods.)

Tony H.


Re: micro-emacs for MVS

Dave Wade
 

-----Original Message-----
From: hercules-380@groups.io <hercules-380@groups.io> On Behalf Of Paul
Edwards
Sent: 15 December 2021 09:05
To: hercules-380@groups.io
Subject: Re: [hercules-380] micro-emacs for MVS

On Wed, Dec 15, 2021 at 06:36 PM, Dave Wade wrote:

1. I need z/PDOS to read/write a tape instead of a 1052+ (which uses
write no cr).
What was the problema with the 1052+
There is no problem with my hacked 1052 except for the fact that under
z/PDOS I basically have BSAM support for a 1052, and that BSAM support
does not exist on MVS, so Hecules/380 allows an EBCDIC ANSI terminal to run
on z/PDOS but not on MVS, and I'd like a solution that works on both
environments.
So you must have some channel programs that work with this. Why don't you tweak these to work with EXCP and wrap it up in a BSAM like layer

BFN. Paul.
Dave


Re: micro-emacs for MVS

Dave Wade
 

-----Original Message-----
From: hercules-380@groups.io <hercules-380@groups.io> On Behalf Of Paul
Edwards
Sent: 15 December 2021 09:05
To: hercules-380@groups.io
Subject: Re: [hercules-380] micro-emacs for MVS

On Wed, Dec 15, 2021 at 06:36 PM, Dave Wade wrote:

1. I need z/PDOS to read/write a tape instead of a 1052+ (which uses
write no cr).
What was the problema with the 1052+
No I think you would have to use EXCP. Doesn't BSAM in MVS 3.8J support any terminal devices?


There is no problem with my hacked 1052 except for the fact that under
z/PDOS I basically have BSAM support for a 1052, and that BSAM support
does not exist on MVS, so Hecules/380 allows an EBCDIC ANSI terminal to run
on z/PDOS but not on MVS, and I'd like a solution that works on both
environments.

BFN. Paul.
Dave


Re: micro-emacs for MVS

Paul Edwards
 

On Wed, Dec 15, 2021 at 06:36 PM, Dave Wade wrote:

1. I need z/PDOS to read/write a tape instead of a 1052+ (which uses write
no cr).
What was the problema with the 1052+
There is no problem with my hacked 1052 except for the
fact that under z/PDOS I basically have BSAM support
for a 1052, and that BSAM support does not exist on
MVS, so Hecules/380 allows an EBCDIC ANSI terminal
to run on z/PDOS but not on MVS, and I'd like a solution
that works on both environments.

BFN. Paul.


Re: micro-emacs for MVS

Dave Wade
 

-----Original Message-----
From: hercules-380@groups.io <hercules-380@groups.io> On Behalf Of Paul
Edwards
Sent: 15 December 2021 00:37
To: hercules-380@groups.io
Subject: Re: [hercules-380] micro-emacs for MVS

On Wed, Dec 15, 2021 at 09:42 AM, Joe Monk wrote:

I would just use TCPIP directly in the program. I think EZA Socket
is still supported.
No, I don't want TCP/IP. I'd rather drive a tape drive than that.
Yall are in la-la land...
Why do you say that? If I am running z/PDOS on z/Arch hardware, there is no
TCP/IP - I haven't written it. But someone can manufacture an EBCDIC ANSI
terminal.
z/PDOS supports BSAM so mvssupa.asm will continue to work.

The only bit that isn't already written is:

1. I need z/PDOS to read/write a tape instead of a 1052+ (which uses write no
cr).
What was the problema with the 1052+

2. Hercules/380 needs to be able to attach a telnet connection to a tape
drive.

BFN. Paul.


Re: micro-emacs for MVS

Paul Edwards
 

On Wed, Dec 15, 2021 at 09:42 AM, Joe Monk wrote:

I would just use TCPIP directly in the program. I think EZA Socket is
still supported.
No, I don't want TCP/IP. I'd rather drive a tape drive
than that.
Yall are in la-la land...
Why do you say that? If I am running z/PDOS on z/Arch
hardware, there is no TCP/IP - I haven't written it. But
someone can manufacture an EBCDIC ANSI terminal.
z/PDOS supports BSAM so mvssupa.asm will continue
to work.

The only bit that isn't already written is:

1. I need z/PDOS to read/write a tape instead of a 1052+
(which uses write no cr).

2. Hercules/380 needs to be able to attach a telnet
connection to a tape drive.

BFN. Paul.


Re: micro-emacs for MVS

Joe Monk
 

Yall are in la-la land...

Joe

On Tue, Dec 14, 2021 at 4:56 AM Paul Edwards <mutazilah@...> wrote:
On Tue, Dec 14, 2021 at 09:50 PM, Dave Wade wrote:

> I would just use TCPIP directly in the program. I think EZA Socket is still
> supported.

No, I don't want TCP/IP. I'd rather drive a tape drive
than that.

BFN. Paul.






Re: micro-emacs for MVS

Paul Edwards
 

On Tue, Dec 14, 2021 at 09:50 PM, Dave Wade wrote:

I would just use TCPIP directly in the program. I think EZA Socket is still
supported.
No, I don't want TCP/IP. I'd rather drive a tape drive
than that.

BFN. Paul.


Re: micro-emacs for MVS

Dave Wade
 

-----Original Message-----
From: hercules-380@groups.io <hercules-380@groups.io> On Behalf Of Paul
Edwards
Sent: 14 December 2021 10:06
To: hercules-380@groups.io
Subject: Re: [hercules-380] micro-emacs for MVS

On Tue, Dec 14, 2021 at 07:55 PM, Dave Wade wrote:

Is treating a tape drive this way odd. As I said it has a historic president.
The precedent you cited was not a S/3X0:

Whilst not a 360, I know the Manchester Medical school had a PDP-8
connected to a IBM7090 as tape drives.
No but same sort of machine. I think the 7090 was the first machine to have channels.


It would be real cool if one day someone built a real EBCDIC ANSI terminal. I
have micro-emacs ready to go. But how will they connect the terminal to a
S/390 or z/Arch machine?

I guess that would be a standard channel connector so it doesn't make any
difference whether it is defined in MVS as a tape drive or anything else. The
physical hardware will be identical, right?
Well modern machines no longer have bus-and tag channels.
Not sure about fibre.
I would just use TCPIP directly in the program. I think EZA Socket is still supported.
Last time I used a real "Z" box was when I tried one of the IBM courses.
You don't get to see the "Z" box, it all runs in JAVA so you have an ISPF like tool on your PC in Java.
You can submit JCL and read the output etc.


Thanks. Paul.

Dave

21 - 40 of 787