Gosh, do I remember those 3278-2 screens. I remember the first time I saw a 3279 4-color :). And then, as the senior systems guy, I got a 3270 PC XT, which emulated four 3270 screens. Thought I was in heaven :)

ESA survivor here. It's amazing to see these panels again.

My father's last work at IBM before he took the early retirement offer in 1985 (he was an SE) was exactly that, helping clients move from MVS to MVS/XA. And at least by what he told me, it was actually kind of a rare skillset, so IBM Germany was left with just one of the three persons with that skillset after the early retirement thing - in the 1980s IBM was for the first time struggling financially, and had to reduce their workforce, and they still believed in not firing anybody, so they offered those in the right age group (55+) early retirement. And since they did it globally, that lead to situations like in Germany where they suddenly didn't have people to do this important work. As for the 3083 I had access to as a IT student in the mid- to late eighties, no idea what was running on it. We did use MVS though, doing a bit of assembly programming, but how much memory do you really need for a proof-of-concept assembly program (don't remember what it was, it was either prime numbers or fibonacci numbers, and not many of them either).

Great stuff, I started my mainframe career in 1986 with … MVS/XA. Traveled back in time, thanks a lot 🙏🏻

"Luxury" My brief experience with the 360/370 line was in the early 1970s. On it the program editing tool I used was not as nice as what you showed. Basically it was about like the MSDOS edline. You needed to give it the line number you wanted to work on and then the command for what to do. Communication was going on at some number for not very many baud.

Super cool video - thanks for putting this together! Great to see a walk through of some MVS assembler (something I never got anywhere near).

I used MVS/XA from 1988 to 1990 or so when the systems went to MVS/ESA and I used that until 1992 or so. There were three mainframes at the installation: a 3090, a 3084, and a 4381. I started on the 3084, had to transition to the 4381 (which was slower), then the 3090 later on.

This is great. I've learning so much about mainframes by watching your videos. I've never had a chance to do that before. They've always seemed very mysterious to me. I'm still waiting for the Pascal or BASIC command line video. :D

"Jagshemash! In Kazakhstan, we not only find potassium and superior gypsy tears, but now, world, behold! We recover the greatest Kazakh treasure, the MVS/XA operating system, M242! Is nice! This system, like our women, very powerful and mysterious, yes? Western countries have Silicon Valley, but we in Kazakhstan, we have Silicon Steppe! High five!” - Borat probably

Our test system in our bank is called XAT. XA-test. I still have to explain it to the newbie’s

Ah, once again, I enjoyed seeing your presentations. :) Interesting, return code of 4, from the assembly, but according to the assembler listing no statements were flagged. Ah, I see, you got a warning (message # IEV140) at the very bottom. Didn't include the 'END' card. Yes, I figured, the assembler module was for demonstration purposes, but I noticed the GETMAIN macro calls which did not utilized the acquired storage. However, it was great review of an assembler language program.

Very nice, great post as always!

Realy Amazing, i like a Lot this design, like a muscle car in a old fashion way

Hehe - time travel or DID IBM produce MVS/XA in 1923 ? That would really turn the last 100 years of history UPSIDE-DOWN. Food for though for someone like Harry Turtledove.

64-bit physical address- that's 16 exbibytes which can be directly addressed. Back in 2004 or so when z/Architecture was introduced this was an almost unfathomable address space. I don't think even most supercomputers had such a large physical address space.

thanks for some nostalgia. I haven't seen ISPF 2.3 (?) in decades, though I used to have a copy then lost it. Anyway, when I first added syntax coloring to ISPF (that later became the HILITE command), I think I did it on ISPF 2.3 as some sort of an exit. If I recall correctly (maybe not), I distributed the HILITE stuff as freeware on a personal website, with IBM's permission, and worked the bugs out so I could eventually slip it into ISPF which was somewhere in V3 by then. same with ISRDDN/DDLIST. that was pretty unusual in those days. Ah, fun times.

Wonderful seeing the simple SPF. Curious if you ever used the SCRIPT/VS program as seen in the Foreground Selection Panel?

Very interesting video... The MOSHIX UTLS have been there before or did you install them?

Veeery nice! - Kazakhstan, Kazakhstan, greatest country in the world, all other countries are ruled by little girls!

Schlumberger might have been the only company operating in Kazahkstan who could afford a mainframe.

Is it running on real hardware, or emulator? How to get access?

Hello, I am looking for a pdf or information that relates a PANEL where there are input fields, with the REXX that takes that data written in the panel and manages it with a jcl. The entire life cycle. For some reason, I only find or panel design or some useless rexx that doesn't work. Do you know about something like that? How do I relate an input field to the rexx! it's not complicated. I can't find anything like that, not even videos. greetings.

Nice thumbnail too!

Before I start describing my concern, here's a quick quip: What is the quickest way to get a retired MVS programmer what is a browser cookie? Ans.: It is analogous to an ISPF profile! Just as insufficient memory gives rise to paging thrashing, I believe that the base plus displacement form of IBM mainframe instruction set addressing design can lead to address reference thrashing, in which so much of memory is eaten up by the constant reloading of "*.USING" address constants! If a provision for indirect addressing with pre and post indexing options that other mainframes had, like the GE600 series machines had would virtually eliminate this issue. I would be interested in how the 64-bit address capabilities handle this issue. My mathematical abilities in dealing with "higher hierogliphics' s limited in constructing a proof. Also, I would be interested in learning if the glaring omission of instructions for save/load multiple for the floating-point registers have been corrected. The issues of addressability needs to be solvled if the huge address space of quantum computers can be dealt with!

hi moshix, can you give me a hint, please? i modified the tk5-included prime number cobol JCL, but when i SUB it the MVS prints that the job is submitted, but i can't find in OUTLIST. obviously i made a mistake in some JCL cards, but i don't know how to get the "log" of the job, at least to understand where i made the mistake. i used CLASS=A and MSGCLASS=H, and i tried ST *, but it seems that i can't see the "failed" jobs. if i run the original job (modifyng only msgclass and job name) it appears in the list, and for certain types of error i can see the failed job (i obtained a IEF452I error, for example). I red some redbooks and mvs-related pdf and i order some JCL books, but i can't undestand how to read the outputs of failed jobs... can you help me or tell me where to get those info? thank you very much

16MB ? come on ! 640KB should be enough for everybody ;-)

Hah, awesome!

Is it possible to make a copy of the os?