This talk about MVS 3.8j brings me right back to my first operator job. I had bought an Apple II Plus a couple years before, and i was such a smarty-pants, I knew "Everything there was" to know about computers, didn't I? Then i got a night operator job, at a lonely RJE station in a textile mill. Just me, and a printer, and a card reader, and a couple of terminals, and i thought "wow, they discovered how to make computers, boring," I was "hot-stuff" because I was 20 and knew so much about the 6502 and how much you could do with it. Well, one night, I was browsing the IT library, and i found a book called "System/370 Principles of Operations"... ...and it was like Neo, waking up in the Power Plant. I was speechless. I never went back, for the next 30 years.
Heck, yes! I first got captivated by computers using the IBM PC. I'd previously taken a couple intro coding classes on the mainframe, using punchards or RJE from an HP minicomputer. But it just didn't do it for me, you know? That all changed when I found the Blue Brown JCL book, and discovered that I could use standard IBM utilities to access files that weren't available to me from the student timesharing system's user interface at school. Oh, I thought I had found the magic key to the treasure chest! "Look, Al, I can see what files you have on your account and i can even copy one. ANYONE can do it! You just have to know the trick. Here's the trick...". The allure of amazing things being possible if you just did the research and knew how to do them had me hooked. The more you learned, the more you could do! How great does that sound? And then, I took an Assembler class and learned that I could use assembler to interrogate the OS system tables in the virtual address space of my program. Suddenly, here was a whole world I could explore, and I came face to face with the sheer scale and complexity of MVS architecture, compared to single user PC systems... all I needed were the maps that told me what I was seeing, the places I could go, and how to get from A to B. I could learn about how the system accommodated so many users, about how it handled I/O, about how jobs were prioritised, etc. -- and monitor it all in close to real time. I read through my university's system reference manuals.... wrote code... got in trouble more than once for writing code ("How do you know how to do that?" "I read the manuals.") ... was a bit careful writing the next set of code... apologised for accidentally breaking ACF2 via its interface to OBS Wylbur (yes, I was the WYL-REC girl)... got more serious and started asking permission before writing code just in case... and the rest was history. I learned so much, and had a great time doing so -- but never got to use it because no one in my area was interested in hiring a 24 year old woman just out of school as an IBM systems programmer, and I wasn't interested in doing my time in COBOL business transaction processing for 4 or 5 years in the hopes that I'd eventually, *maybe* have the chance to move into systems. The "maybe" part was the deal-killer. I went to my second choice option, UNIX. That doesn't mean that I haven't occasionally pulled out the MVS knowledge I've retained for decades as a bit of a party trick to surprise people who have no idea that before Random was a UNIX TCP/IP networking programmer, she was a relatively hardcore MVS systems enthusiast. Nearly 40 years later I can say pretty authoritatively, I will never know another assembly language as well as I knew IBM assembler.
The reaseon MVS 3.8 is in the public domain is because the federal government paid for the research. In the case of the IBM antitrust lawsuit, there was no settlement. The US government dropped the case because it was too big to litigate and apparently IBM was likely to win. IBM, however, decided during the antitrust litigation, (probably to weaken the government's case) to unbundle its operating systems from its computers, which, while it meant you didn't have to buy IBM hardware to be able to use their software, most customers saw a net increase in costs compared to the old bundled sale system.
Congratulations Moshix for your extraordinary capability to produces this summary and we could said that remember is to be alive ...in parallel were developed from others operating systems which were by Germany (Happily meritorious in Boeblingen-Sindelfingen) as DOS/SP, DOS/VSE, SSX/VSE, z/VSE etc
Thanks for your comment - it saved me from submitting my own... I worked in the DOS/VS/ESA arena for over 40 years and loved its simplicity. DOS/360 actually precedes OS/360 by about a year. At one point in time there were more DOS/VS licenses than any other 370 operating system. I wouldn't mind having a DOS/VS play ground system but the problem is since IBM didn't provide a lot of supporting software with the operating system, having a plain vanilla DOS/VS system wouldn't be too useful.
Back in the 1980s, memory was about USD $1 per K. Now, it wasn't for IBM machines, but for comparison, Digital Equipment Corporation sold 100 MB disk drives that cost $27,000. Control Data Corporation sold similar drives for $7,000, but they required a $300 controller card. Additional 100 MB removable disk packs cost $700. So 16 MB of memory then, cost about $16,000.
I seem to recall the cost of 1 MiB of RAM in 1983 being about $800, for the Lisa PC. By 1997/1998, the cost of RAM was approximately $800-$900 for 1,024 MiB, or 1 GiB.
Thank you Moshix!!! Very nice good video. Answered to all my questions and apparently I wasn't the only one having these questions in mind. For me, you've said it and I'm stopping chasing for ghosts and run zOS just because it's more like into 2020. I'm a newby anyways but thank's to you, after 20 years in IT (developer/sys admin/net admin/CISO) I'm trying to steer the wheels turn to mainframes . I love IBM philosophy and worked with AIX, PowerVM, Storwize and Power 7 series hardware.. but I like mainframes for whatever reason. Maybe because it relies on a simple terminal.
Update on tools used in z/OS, include running SAS jobs, SAS is used extensively by business/risk users trying to make sense of large datasets (many millions of rows/transactions). SAS makes it manageable to do validation, and analysis of both UAT and production data.
3.8 on a big (9 meg!!) Amdahl V8 is what i operated on as an RJE tenant user before i went sysprog and brought MVS/SP into our own company, on a 8MB 4381. The one thing i miss, from the 3.8 days, is ROSCOE (at that time an ADR product.) I wish CA would allow hobby licenses of that fine product. Filesystems: I would definitely think, that what the OS provides us, through DADSM and the various access methods, counts as a file-system! Maybe it's not sophisticated enough for some people (though VSAM definitely is sophisticated, it's not simple!!!) but I say, it all counts! Of course modern USS does run a filesystem designed to mount/span multiple volumes, but you can still manage files, and very well, spanning/mixing both DASD and tape media. Maybe it's a "collection" of file-systems? security: You could also create a "password" file on each disk volume (we didn't have RACF for some years,) and set a bit in the VTOC entry for a disk file, marking it as "password-protected". it's pretty crude, if i remember, a TSO user trying to open a password-protected file, might even cause a write-to-console-operator, requesting the password...a batch job, definitely would. It's pretty laughable (and if you know your way around the "superzap" utility, you can break any security,) but it is definitely password security. you can use it to lock down TSO logins by locking down sys1.uads and sys1.brodcast - i think - at least you could, under MVS/SP which was not very different from 3.8.
Exactly. I was going to make the same comment. Individual files could be passworded. A TSO user attempting to open a passworded file would get a prompt asking for the password. On a batch job, the console operator would get the request for the password. The password had to be included in the operations documentation OR - if I recall - fed to it by some utility program with the password presented in clear text in the JCL. Sure awkward and insecure. Back in the day, ACF2 was the overwhelming choice over RACF which itself was horrendous to implement. ACF2 came first, being a commercial implementation of a security system developed at the University of Illinois at Urbana-Champaign. SKK were the initials of the last names of the developers/founders. RACF did improve later. Much later. And certainly, one of the funnest things about MVS is being able to play with Superzap and all the other naughty system utilities.
@@KameraShy I totally forgot about ACF2. The service bureau running the Amdahl used ACF2, under 3.8j. That was a pretty cool system. We cheaped out, when we went in-house on SP, and we used the password files. We also didn't have TMS, so we had a lot of "fun" and some scary moments, doing scratch-tape management "the hard way". Thank God for tape rings, is all i can say. "WHAT, is the LAW? Tapes to the tape-rack, Rings to the ring-box, Sir!" ;)
Another great video, Thanks! I'm an old programmer, loved COBOL in college and at my first job running on a PR1ME Minicomputer. But lately have been interested in learning the "Mainframe" environment and this video helped a lot!! :) BTW, whatever happened to "Mini"Computers? It looks like they've become extinct except maybe for hobbyist reliving their earlier careers?
@@moshixmainframechannel Yes, but to clarify, by extinct I mean Minicomputers like the Pr1me aren't commonly used commercially if at all, right? I find learning the Mainframe environment more exciting because it's still used commercially and it's still possible to get paying jobs writing/maintaining COBOL on them. Not that I expect to find a job, but the possibilities are like carrot on a stick to keep me interested and motivated.
Hi Moshix, Thank you for putting these videos together. I am new to MVS and trying to grasp the basics. I'm looking for a 101/in a nutshell kind of resource that will allow me to spend a quick hour here and there getting to grips with the system. Do you have one or two videos in your series that cater for that or is there a quck reference guide? Sorry if these are obvious or already covered questions, I'm a litle time poor just now so I'm looking for a way to maximise my enjoyment in the first steps of a learning curve.
I've played with z/OS before, but for any sort of project that I would want to share with other people, I would definitely use MVS 3.8 instead. It would be a shame to put hard work into something and then have it be shut down due to licensing.
Dear Moshix, I fully endorse your view and verdict: MVS 3.8j is a lot of fun to use and explorer, especially when you're new to the mainframe or have been bound to MVS userland before (that is, TSO with limited capabilities + permissions). Not messing with any company's lawyers that for sure have a strong position in what they defend is also good advice. Personally, I would not have any objection if somebody does run z/whatever illegally on Hercules, at the end of the day people are usually grown up adults and should know what they're doing. About licensing policy and anti trust, because you've mentioned. I'm also surprised that IBM and - a similar case I daresay with Apple - can get away with binding the software to the hardware. I know, I could write - in theory - Gromi/OS and run it on IBM hardware but not z/OS on my hardware, be it real or emulated. What was running on the Amdahl machines back then? Or today? Anyway, you're message is clear and as I said, endorse your views and would like to encourage everybody who's interested in Mainframe concepts not only to start with MVS 3.8j, but also embrace it. Soon one will start to realise that it's not only an "abandoned" free-by-ant-trust-law thing as sexy as MS-DOS 1.0, but a very capable piece of software backed and maintained by a group of excellent contributors. You've mentioned Jürgen Winkelmann but there are many others. It's so exciting!
Other Equipment Manufacturers could build hardware that ran MVS, and there were multiple companies that did, but all gave up when IBM went 64 bit. I have worked om Fujitsu and Amdahl equipment with MVS.
@@brucehewson5773 I worked on the IBM side.... although I was on the As/400, the mainframe support was next. I witness too many times blaming one side or the other if something went wrong. Was the same for me on the AS/400 when some 3rd parties cards or equipment was connected to the AS/400... we were instructed to say "you have to reproduce it on IBM hardware" ... :-/
@@gippa75 Before AS/400 were the Series 34 & 36 machines. Early 80's I was tasked with getting them connected to MVS. RJE support = remote job access to JES2. Took a few tweaks of settings but it worked. At the same time, small systems abounded. Also tasked with getting each and every one, no matter the communication setup, to get their connections working. Was a fun time of growth and exposure to things outside on the mainframe I had been working on. Large system also, connection to a FACOM machine, pretty sure we got NJE working on that connection.
As far as patents are concerned, unless they were granted after 2002, they have expired. So anything patented in IBM hardware made before 2002 is now public domain. Same for any software patents (which aren't even allowed in some countries). Now, it's entirely possible that IBM requiring ownership of one of their computers in order to purchase their software, that might be a "tying contract" which violates antitrust law. The only reason for all of this is so that IBM can bill people much more than they could when customers owned these things. It's a damn shame.
Parents expire after 20 years, correct. As for the tying contract. They got that confirmed by a U.S. judge. The fear is not about being wrong. The fear is about spending money to mount a defense
Hi Moshix. Thanks for your insights. They are very useful when you are debuting in the mainframe world. I am curious to know how would MVS 3.8 tk4 cater for the training needed by an application programmer who intend to work in modern mainframe system. Is there a way to install in MVS 3.8 a COBOL 85 compiler? Regarding DB2 if I am not mistaken, there is no way to make it work on MVS due to technical limitations inherent to the OS. Regarding CICS, there seems to be an alternative as you depicted in a previous video. Apart from that, I honestly do not know how an application programmer can be trained outside the Z/OS ecosystem using exclusively MVS 3.8 TK4. I hope it could be possible, please correct me if I am mistaken. Thank you.
Hi Moshix. Could you please elaborate on the state of update 9 and TCP/IP? The only thing I've found is cbttape.org/~jmorrison/mvs38j-ip/index.html and it's not very auspicious, it hasn't been updated in years. I believe the obsession with z/OS might stem from the fact that we can do little networking with mvs38.
You did not mention OS/VS2 SVS (single virtual system). Do you know if there were shops using it? What about MFT? Was it ever used in production? Do we have the tapes to sysgen one?
I didn’t mention all the steps in the lineage because I was focusing on how MVS became z/OS. There were other steps I left out. MFT was used extensively by American Airlines for example. And many European customers also. The tapes for genning MFT are still around.
Still I find this emulation rather interesting. So, operating system - MVS 3.8 is restricted 16mb. Yet, I'm assuming the currently hardware emulation is capable of much more, but is restricted by the limitations of the operating system. I suppose, if I really don't like any of the available compilers, I could write my own in assembler.
The limitation is not by the operating system but rather by the architecture which only has 24 bits for addressing memory. You can of course write your own assembler
Funny that people get confused about what OS is what with IBM... Funny in the sense that I think IBM has always purposely confused people with terminology as a marketing strategy. IBM has always been it's own little world of terminology. At some point I should do a tongue in cheek translation dictionary between "regular computer terminology" to IBM terminology.
There was a substantial difference. I was at the IBM AS/400 L3 Support and I didn't know how to run S/36 and S/38. they were similar in concept, but different. As a matter of fact I can remember the S/36 and S/38 subsystem for AS/400. Side node for @moshix.... I was obsessed for a certain period of time for an As/400 emulator, but I believe there was no free version of the OS. I really loved OS/400, it was very popular in my home country (Italy) as it could range from a small desktop to a near-to-mainframe version.
@@gippa75 When I worked in ops we had AS/400s as well as Amdahld (and, subsequently, IBM) mainframes. I really liked the AS/400, it was an interesting approach by IBM and the 'everything in the box' approach, as well as it being object oriented, was really cool. As an operator I preferred MVS though. That said I'd love to see an AS/400 / OS/400 emulator but I don't think we'll ever see that happen.
@@theimp67 I tried, but there are too many things implemented at microcode level. We're talking about years of work :( And I gave up. (not mentioning the legal issues)
@@theimp67 Although I'm pretty late, I'm pretty sure there's a as400 emulator on AWS. You can connect a 5270(?) to it and play. I think you could play quite a bit under the free usage time.
One 👎??? What are they, a Unixcorn!?! Will this run adequately on Windows System 10 or should one install Linux on my Dell Laptop to play around with it??? My assumption is that this can only allow you to code COBOL in ISPF, compile and run it with QSAM or VSAM? ie no DB2 or IMS?
Naw, there are no “racists” in IT (only Techies who think there is only one way to code and one language to code in... languagists or “linguists” for short 🤣😅😂
Hi Moshix, When I press F7 button to scroll down the dataset it is coming out abruptly and not allowing me to login for a while saying..already user loggin..Can you help me with this issue Please..This is a really fantastic job and helping us a lot.. thanks again
A great history of the IBM mainframe systems from System 360 and beyond: groups.io/g/hercules-os380/files/_$OSTL37.5.pdf You may need an account with groups.io to get it. :)
Yes I know it. And I have the PL/1 source code. However the code only compiled with Optimizer PLI compiler which we have but doesn’t work reliably on MVS 3.8. One could compile it on OS/390 and try to make it work on MVS 3.8 but it will probably get sick from instructions not present in S/370
I have been Interested in old computers from the 1950s to the early 1990s for quite some time. Recently, I've been interested in the DEC PDP-10 and IBM Mainframe. I would like to get the latest version of Tops-10 running on the Simh KL10 emulator and play around with the software that was available for that operating system. I'll have to watch your VAX video when it comes out.
Listen, nobody would want to run IBM operating systems on a PDP 10 or VAX. Having used a VAX, a Decsyatem 20, and MVS, I can tell you MVS as OS JCL is worse than anything Digital produced. People used IBM software because it was there and no competitor, even if their offering was better, could get people to switch. Nobody like MVS; they used it because they had to. IBM was legendary in creating "fear, uncertainty, and doubt" (FUD) about competitor offerings, which were often technologically superior, but IBM had the marketing dollars to make others look worse. As the saying was, "nobody ever got fired for buying IBM."
Hi Moshix, I signed up to watch your Vax 11 video you made that was available at 10 am the other day. I slept in (I'm retired) and I didn't get to see it. Later I tried to view it and was told it was private. How can I view it?
This talk about MVS 3.8j brings me right back to my first operator job. I had bought an Apple II Plus a couple years before, and i was such a smarty-pants, I knew "Everything there was" to know about computers, didn't I? Then i got a night operator job, at a lonely RJE station in a textile mill. Just me, and a printer, and a card reader, and a couple of terminals, and i thought "wow, they discovered how to make computers, boring," I was "hot-stuff" because I was 20 and knew so much about the 6502 and how much you could do with it. Well, one night, I was browsing the IT library, and i found a book called "System/370 Principles of Operations"...
...and it was like Neo, waking up in the Power Plant. I was speechless.
I never went back, for the next 30 years.
Heck, yes! I first got captivated by computers using the IBM PC. I'd previously taken a couple intro coding classes on the mainframe, using punchards or RJE from an HP minicomputer. But it just didn't do it for me, you know?
That all changed when I found the Blue Brown JCL book, and discovered that I could use standard IBM utilities to access files that weren't available to me from the student timesharing system's user interface at school. Oh, I thought I had found the magic key to the treasure chest! "Look, Al, I can see what files you have on your account and i can even copy one. ANYONE can do it! You just have to know the trick. Here's the trick...". The allure of amazing things being possible if you just did the research and knew how to do them had me hooked. The more you learned, the more you could do! How great does that sound?
And then, I took an Assembler class and learned that I could use assembler to interrogate the OS system tables in the virtual address space of my program. Suddenly, here was a whole world I could explore, and I came face to face with the sheer scale and complexity of MVS architecture, compared to single user PC systems... all I needed were the maps that told me what I was seeing, the places I could go, and how to get from A to B. I could learn about how the system accommodated so many users, about how it handled I/O, about how jobs were prioritised, etc. -- and monitor it all in close to real time. I read through my university's system reference manuals.... wrote code... got in trouble more than once for writing code ("How do you know how to do that?" "I read the manuals.") ... was a bit careful writing the next set of code... apologised for accidentally breaking ACF2 via its interface to OBS Wylbur (yes, I was the WYL-REC girl)... got more serious and started asking permission before writing code just in case... and the rest was history.
I learned so much, and had a great time doing so -- but never got to use it because no one in my area was interested in hiring a 24 year old woman just out of school as an IBM systems programmer, and I wasn't interested in doing my time in COBOL business transaction processing for 4 or 5 years in the hopes that I'd eventually, *maybe* have the chance to move into systems. The "maybe" part was the deal-killer. I went to my second choice option, UNIX. That doesn't mean that I haven't occasionally pulled out the MVS knowledge I've retained for decades as a bit of a party trick to surprise people who have no idea that before Random was a UNIX TCP/IP networking programmer, she was a relatively hardcore MVS systems enthusiast. Nearly 40 years later I can say pretty authoritatively, I will never know another assembly language as well as I knew IBM assembler.
Wow. Thanks for sharing !
Moshix's knowledge and contributions worth tons of gold. Thanks for sharing!
Oh thank you
The reaseon MVS 3.8 is in the public domain is because the federal government paid for the research. In the case of the IBM antitrust lawsuit, there was no settlement. The US government dropped the case because it was too big to litigate and apparently IBM was likely to win. IBM, however, decided during the antitrust litigation, (probably to weaken the government's case) to unbundle its operating systems from its computers, which, while it meant you didn't have to buy IBM hardware to be able to use their software, most customers saw a net increase in costs compared to the old bundled sale system.
No. In this case you are wrong. It’s a myth that it is open source because the treasury paid for parts of it.
Congratulations Moshix for your extraordinary capability to produces this summary and we could said that remember is to be alive ...in parallel were developed from others operating systems which were by Germany (Happily meritorious in Boeblingen-Sindelfingen) as DOS/SP, DOS/VSE, SSX/VSE, z/VSE etc
I seem to remember DOBIS/LIBIS was developed there
Thanks for your comment - it saved me from submitting my own... I worked in the DOS/VS/ESA arena for over 40 years and loved its simplicity. DOS/360 actually precedes OS/360 by about a year. At one point in time there were more DOS/VS licenses than any other 370 operating system. I wouldn't mind having a DOS/VS play ground system but the problem is since IBM didn't provide a lot of supporting software with the operating system, having a plain vanilla DOS/VS system wouldn't be too useful.
Thanks Moshix. Another great video. Really, MVS 3.8j is a great play ground to learn all about mainframes.
Back in the 1980s, memory was about USD $1 per K. Now, it wasn't for IBM machines, but for comparison, Digital Equipment Corporation sold 100 MB disk drives that cost $27,000. Control Data Corporation sold similar drives for $7,000, but they required a $300 controller card. Additional 100 MB removable disk packs cost $700. So 16 MB of memory then, cost about $16,000.
I seem to recall the cost of 1 MiB of RAM in 1983 being about $800, for the Lisa PC. By 1997/1998, the cost of RAM was approximately $800-$900 for 1,024 MiB, or 1 GiB.
Thank you Moshix!!! Very nice good video. Answered to all my questions and apparently I wasn't the only one having these questions in mind. For me, you've said it and I'm stopping chasing for ghosts and run zOS just because it's more like into 2020. I'm a newby anyways but thank's to you, after 20 years in IT (developer/sys admin/net admin/CISO) I'm trying to steer the wheels turn to mainframes . I love IBM philosophy and worked with AIX, PowerVM, Storwize and Power 7 series hardware.. but I like mainframes for whatever reason. Maybe because it relies on a simple terminal.
Update on tools used in z/OS, include running SAS jobs, SAS is used extensively by business/risk users trying to make sense of large datasets (many millions of rows/transactions). SAS makes it manageable to do validation, and analysis of both UAT and production data.
3.8 on a big (9 meg!!) Amdahl V8 is what i operated on as an RJE tenant user before i went sysprog and brought MVS/SP into our own company, on a 8MB 4381.
The one thing i miss, from the 3.8 days, is ROSCOE (at that time an ADR product.) I wish CA would allow hobby licenses of that fine product.
Filesystems: I would definitely think, that what the OS provides us, through DADSM and the various access methods, counts as a file-system! Maybe it's not sophisticated enough for some people (though VSAM definitely is sophisticated, it's not simple!!!) but I say, it all counts! Of course modern USS does run a filesystem designed to mount/span multiple volumes, but you can still manage files, and very well, spanning/mixing both DASD and tape media. Maybe it's a "collection" of file-systems?
security: You could also create a "password" file on each disk volume (we didn't have RACF for some years,) and set a bit in the VTOC entry for a disk file, marking it as "password-protected". it's pretty crude, if i remember, a TSO user trying to open a password-protected file, might even cause a write-to-console-operator, requesting the password...a batch job, definitely would. It's pretty laughable (and if you know your way around the "superzap" utility, you can break any security,) but it is definitely password security. you can use it to lock down TSO logins by locking down sys1.uads and sys1.brodcast - i think - at least you could, under MVS/SP which was not very different from 3.8.
Exactly. I was going to make the same comment. Individual files could be passworded. A TSO user attempting to open a passworded file would get a prompt asking for the password. On a batch job, the console operator would get the request for the password. The password had to be included in the operations documentation OR - if I recall - fed to it by some utility program with the password presented in clear text in the JCL. Sure awkward and insecure. Back in the day, ACF2 was the overwhelming choice over RACF which itself was horrendous to implement. ACF2 came first, being a commercial implementation of a security system developed at the University of Illinois at Urbana-Champaign. SKK were the initials of the last names of the developers/founders. RACF did improve later. Much later. And certainly, one of the funnest things about MVS is being able to play with Superzap and all the other naughty system utilities.
@@KameraShy I totally forgot about ACF2. The service bureau running the Amdahl used ACF2, under 3.8j. That was a pretty cool system. We cheaped out, when we went in-house on SP, and we used the password files. We also didn't have TMS, so we had a lot of "fun" and some scary moments, doing scratch-tape management "the hard way". Thank God for tape rings, is all i can say. "WHAT, is the LAW? Tapes to the tape-rack, Rings to the ring-box, Sir!" ;)
Yes, the ES/9000 -- I still think thatt was a tip of the hat to Kubrick fans....
the learner's edition of z/os is 300GB for the ADCD files alone. mvs 3.8j w/ Rob Prins mods are about 600MB.
True
300 gb? does it contains cics, db2, ims, and all other stuff?
@@massimo79mmm eh si.
Another great video, Thanks! I'm an old programmer, loved COBOL in college and at my first job running on a PR1ME Minicomputer. But lately have been interested in learning the "Mainframe" environment and this video helped a lot!! :)
BTW, whatever happened to "Mini"Computers? It looks like they've become extinct except maybe for hobbyist reliving their earlier careers?
Prime emulators exist as as I can remember
@@moshixmainframechannel Yes, but to clarify, by extinct I mean Minicomputers like the Pr1me aren't commonly used commercially if at all, right? I find learning the Mainframe environment more exciting because it's still used commercially and it's still possible to get paying jobs writing/maintaining COBOL on them. Not that I expect to find a job, but the possibilities are like carrot on a stick to keep me interested and motivated.
@@montpierce424 yes you’re right. They are all gone. The PC ate everything else, including the mainframe, really.
Hi Moshix,
Thank you for putting these videos together.
I am new to MVS and trying to grasp the basics.
I'm looking for a 101/in a nutshell kind of resource that will allow me to spend a quick hour here and there getting to grips with the system.
Do you have one or two videos in your series that cater for that or is there a quck reference guide?
Sorry if these are obvious or already covered questions, I'm a litle time poor just now so I'm looking for a way to maximise my enjoyment in the first steps of a learning curve.
I've played with z/OS before, but for any sort of project that I would want to share with other people, I would definitely use MVS 3.8 instead. It would be a shame to put hard work into something and then have it be shut down due to licensing.
Exactly
Dear Moshix, I fully endorse your view and verdict: MVS 3.8j is a lot of fun to use and explorer, especially when you're new to the mainframe or have been bound to MVS userland before (that is, TSO with limited capabilities + permissions). Not messing with any company's lawyers that for sure have a strong position in what they defend is also good advice. Personally, I would not have any objection if somebody does run z/whatever illegally on Hercules, at the end of the day people are usually grown up adults and should know what they're doing. About licensing policy and anti trust, because you've mentioned. I'm also surprised that IBM and - a similar case I daresay with Apple - can get away with binding the software to the hardware. I know, I could write - in theory - Gromi/OS and run it on IBM hardware but not z/OS on my hardware, be it real or emulated. What was running on the Amdahl machines back then? Or today? Anyway, you're message is clear and as I said, endorse your views and would like to encourage everybody who's interested in Mainframe concepts not only to start with MVS 3.8j, but also embrace it. Soon one will start to realise that it's not only an "abandoned" free-by-ant-trust-law thing as sexy as MS-DOS 1.0, but a very capable piece of software backed and maintained by a group of excellent contributors. You've mentioned Jürgen Winkelmann but there are many others. It's so exciting!
Agree!
Other Equipment Manufacturers could build hardware that ran MVS, and there were multiple companies that did, but all gave up when IBM went 64 bit. I have worked om Fujitsu and Amdahl equipment with MVS.
@@brucehewson5773 I worked on the IBM side.... although I was on the As/400, the mainframe support was next. I witness too many times blaming one side or the other if something went wrong. Was the same for me on the AS/400 when some 3rd parties cards or equipment was connected to the AS/400... we were instructed to say "you have to reproduce it on IBM hardware" ... :-/
@@gippa75 Before AS/400 were the Series 34 & 36 machines. Early 80's I was tasked with getting them connected to MVS. RJE support = remote job access to JES2. Took a few tweaks of settings but it worked. At the same time, small systems abounded. Also tasked with getting each and every one, no matter the communication setup, to get their connections working. Was a fun time of growth and exposure to things outside on the mainframe I had been working on. Large system also, connection to a FACOM machine, pretty sure we got NJE working on that connection.
As far as patents are concerned, unless they were granted after 2002, they have expired. So anything patented in IBM hardware made before 2002 is now public domain. Same for any software patents (which aren't even allowed in some countries). Now, it's entirely possible that IBM requiring ownership of one of their computers in order to purchase their software, that might be a "tying contract" which violates antitrust law. The only reason for all of this is so that IBM can bill people much more than they could when customers owned these things. It's a damn shame.
Parents expire after 20 years, correct. As for the tying contract. They got that confirmed by a U.S. judge. The fear is not about being wrong. The fear is about spending money to mount a defense
Hi Moshix. Thanks for your insights. They are very useful when you are debuting in the mainframe world. I am curious to know how would MVS 3.8 tk4 cater for the training needed by an application programmer who intend to work in modern mainframe system. Is there a way to install in MVS 3.8 a COBOL 85 compiler? Regarding DB2 if I am not mistaken, there is no way to make it work on MVS due to technical limitations inherent to the OS. Regarding CICS, there seems to be an alternative as you depicted in a previous video. Apart from that, I honestly do not know how an application programmer can be trained outside the Z/OS ecosystem using exclusively MVS 3.8 TK4. I hope it could be possible, please correct me if I am mistaken. Thank you.
You could use the IBM master the mainframe program to train on z/OS. Or buy a zPDT license
Hi Moshix. Could you please elaborate on the state of update 9 and TCP/IP? The only thing I've found is cbttape.org/~jmorrison/mvs38j-ip/index.html and it's not very auspicious, it hasn't been updated in years. I believe the obsession with z/OS might stem from the fact that we can do little networking with mvs38.
We will have to wait for Juergen to feel ready to release it. I rather wait and know it’s good than release now and it’s not ready.
You did not mention OS/VS2 SVS (single virtual system). Do you know if there were shops using it? What about MFT? Was it ever used in production? Do we have the tapes to sysgen one?
I didn’t mention all the steps in the lineage because I was focusing on how MVS became z/OS. There were other steps I left out. MFT was used extensively by American Airlines for example. And many European customers also. The tapes for genning MFT are still around.
OS/VS2 was used by some insurance companies, I know. Like Zürich Insurance and Munich Re. For a brief period
Still I find this emulation rather interesting. So, operating system - MVS 3.8 is restricted 16mb. Yet, I'm assuming the currently hardware emulation is capable of much more, but is restricted by the limitations of the operating system. I suppose, if I really don't like any of the available compilers, I could write my own in assembler.
The limitation is not by the operating system but rather by the architecture which only has 24 bits for addressing memory. You can of course write your own assembler
Funny that people get confused about what OS is what with IBM... Funny in the sense that I think IBM has always purposely confused people with terminology as a marketing strategy. IBM has always been it's own little world of terminology.
At some point I should do a tongue in cheek translation dictionary between "regular computer terminology" to IBM terminology.
IBM facelifted the S/38 and called it the AS/400. There was no technical difference between the two at the time, just a name change.
You are certainly wrong about that sir
There was a substantial difference. I was at the IBM AS/400 L3 Support and I didn't know how to run S/36 and S/38. they were similar in concept, but different. As a matter of fact I can remember the S/36 and S/38 subsystem for AS/400. Side node for @moshix.... I was obsessed for a certain period of time for an As/400 emulator, but I believe there was no free version of the OS. I really loved OS/400, it was very popular in my home country (Italy) as it could range from a small desktop to a near-to-mainframe version.
@@gippa75 When I worked in ops we had AS/400s as well as Amdahld (and, subsequently, IBM) mainframes. I really liked the AS/400, it was an interesting approach by IBM and the 'everything in the box' approach, as well as it being object oriented, was really cool. As an operator I preferred MVS though. That said I'd love to see an AS/400 / OS/400 emulator but I don't think we'll ever see that happen.
@@theimp67 I tried, but there are too many things implemented at microcode level. We're talking about years of work :( And I gave up. (not mentioning the legal issues)
@@theimp67 Although I'm pretty late, I'm pretty sure there's a as400 emulator on AWS. You can connect a 5270(?) to it and play. I think you could play quite a bit under the free usage time.
One 👎??? What are they, a Unixcorn!?!
Will this run adequately on Windows System 10 or should one install Linux on my Dell Laptop to play around with it??? My assumption is that this can only allow you to code COBOL in ISPF, compile and run it with QSAM or VSAM? ie no DB2 or IMS?
Yes it will run nice on both Windows and Linux. Why they gave me a thumbs down, I don’t know. May be racially motivated
Naw, there are no “racists” in IT (only Techies who think there is only one way to code and one language to code in... languagists or “linguists” for short 🤣😅😂
What if I want to practice cics+db2? There is no way with mvs 3.8 right?
Correct
Hi Moshix,
When I press F7 button to scroll down the dataset it is coming out abruptly and not allowing me to login for a while saying..already user loggin..Can you help me with this issue Please..This is a really fantastic job and helping us a lot..
thanks again
Don’t press it then? Type down and enter
A great history of the IBM mainframe systems from System 360 and beyond: groups.io/g/hercules-os380/files/_$OSTL37.5.pdf You may need an account with groups.io to get it. :)
Thanks for sharing. Nice document.
если у меня это в рекомендациях, видимо я обреченный гик...
согласен !
Good morning, is there a RDBMS for MVS that can be used legally?
thank you
Nope
@@moshixmainframechannel thank you
@@moshixmainframechannel Hi moshix, i was remembering that actually there is a dbms for mvs.
look here mitrol.info/
Yes I know it. And I have the PL/1 source code. However the code only compiled with Optimizer PLI compiler which we have but doesn’t work reliably on MVS 3.8. One could compile it on OS/390 and try to make it work on MVS 3.8 but it will probably get sick from instructions not present in S/370
Does that mean that if you can port MVS 3.8 to the DEC PDP-10 or VAX, you can legally run it, but that is not the case with Z/OS?
Yes
I have been Interested in old computers from the 1950s to the early 1990s for quite some time. Recently, I've been interested in the DEC PDP-10 and IBM Mainframe. I would like to get the latest version of Tops-10 running on the Simh KL10 emulator and play around with the software that was available for that operating system. I'll have to watch your VAX video when it comes out.
Listen, nobody would want to run IBM operating systems on a PDP 10 or VAX. Having used a VAX, a Decsyatem 20, and MVS, I can tell you MVS as OS JCL is worse than anything Digital produced. People used IBM software because it was there and no competitor, even if their offering was better, could get people to switch. Nobody like MVS; they used it because they had to. IBM was legendary in creating "fear, uncertainty, and doubt" (FUD) about competitor offerings, which were often technologically superior, but IBM had the marketing dollars to make others look worse. As the saying was, "nobody ever got fired for buying IBM."
Hi Moshix. I noticed that the z/OS side has "MVS Acct". That a throwback in the OS or is that your account name?
That’s still in z/os
Hi Moshix, I signed up to watch your Vax 11 video you made that was available at 10 am the other day. I slept in (I'm retired) and I didn't get to see it. Later I tried to view it and was told it was private. How can I view it?
Will be posted on this channel here tomorrow . Thanks for writing Bill !