To clarify, z/OS also has a shell environment like Unix with similar lineage as a timesharing system. It's TSO/E, and it has commands for pretty much all the functionality shown above in ISPF. Shell scripts in this environment are called CLISTS, and they have the normal sort of control flow constructs one might expect. ISPF is sort of like emacs or Midnight commander. It runs on top and provides a pseudo-graphical environment that uses 3270 full screen editing. Because mainframes have more sophisticated terminal protocols compared to Unix, the full screen UI pretty much replaced the line-based shell. It's as if emacs replaced bash entirely and gained TUI support for all system administration tasks. You can also bypass 3270 and ssh into the unix subsystem environment, and you get a pretty vanilla (and somewhat anemic) POSIX shell that behaves as you would expect. You can call out to the TSO/E shell from the POSIX shell.
Great video Moshix. I like that this video combines lessons from your other videos in a nice digestible form. It will go in my quick reference section, for sure!
Hey moshix, thank you very much! This video explains things so nicely for total noobs :) I had some brief experience w IBM back in early '90s (at my university), but everything was totally diff from MVS | Z/OS. That was VM/ESA and we had almost "normal" files (filename 8 ch , filetype 8 ch, filemode 2 [A-Z;0-6]) on minidiscs. We could 'q disk', 'filelist' etc. No dirs, but one could 'define' personal temp minidisc. Datasets/records logic is still totally new to me, and thanks to you I think I'm finally beginning to understand :D
I guess the very first thing most Linux people stumble upon (even before the file system) is the difference of the 3270-terminal logic from the tty-like terminals.
OMG. That looks like pain, only pain. Why would anybody in this universe choose using this postbox like cli-gui instead of any shell with pipeing commands?
I am wondering what pre-interactive-terminal input/output devices the 3270-terminal inherited its IO-metaphor from. Were those punch carded blank forms? Cause this kind of interface is very obviously not implementable for TTY.
@eugrus block I/O is where it comes from. The line length of 80 is from the punch card yes, but the 3270 terminal paradigm comes from the concept of block I/O for efficiency
@@moshixmainframechannel thank you for taking time to answer this! What I kind of meant is: What was the first hardware implementation of the block I/O before interactive terminals? In other words, what was used to interface with OS/360 at the time when Unix was mostly interfaced with via a telex/hardcopy-terminal. Or was an interactive terminal required to connect to OS/360 from the early on?
Watching my as/400 and other IBM counterparts do the constant menu diving for the simplest task always drove me mad. The entire thing always looked less like they were actually doing anything directly in an OS and more like they were just using a text UI “shell” application on top of something IBM was never going to give them real access to. Even programmer options looked like so much hand holding and guarded rails driven through menus. I used to have to modify JCL for jobs back in the early 90s but was just surprised as hell to walk into a shop still using this in 2019.
@@moshixmainframechannel The programmers retrained onto other systems easily enough. However the “power users” who had just memorized the menu sequences (“3, 5, 1”, “1,1,6”) took forever to retrain on mouse driven menus lol.
It's interesting that z/OS requires the user to specify the actual amount of space to allocate for the file (excuse me, the "dataset") before its creation. However, I wonder if that opens up a potential security problem. Many decades ago when I was a university student I had an account on a PDP 11/70 that ran RSTS/E. That OS didn't require it but you could tell it to allocate a specified number of bytes to the file at creation time. However, this feature not only created a file ofy the size you wanted, it left it full of whatever data that was written in the free space allocated to the file! I found it fun to allocate big files and then pop it into TECO and see what non-binary text there was to see. Does z/OS suffer from this or has IBM done something about it, hopefully LONG ago?
List all the files in all the disks: you can’t to that in traditional Windows either. The reason one can in Unix is because all the directories / folders have a common root. There is no C: that van be confusing for Linux Users at the beginning.
Would you consider for new projects on a green field approach any business case still to use a mainframe? I am watching you channel for days in a row and till now I can not see why I should start any project on this outdated concept. Since the introduction of Kubernetes and Micro Services like Netflix, I believe the business case got canceled for any new Mainframe
The physical mainframe machine has many advantages, like hotswappable CPUs and RAM. Mainframes are not needed for common tasks like hosting websites, because those workloads are both easilly distributable over many small machines and non-critical systems.
Hi Moshix I am working a project to basically create a n Oracle VM Virtualbox interface to create linux images on a bare metal mainframe. I need some help. If you would be interested in assisting (paid) please reach out to me. If you know of anyone else that be interested please put me in touch with them. Thanks
To clarify, z/OS also has a shell environment like Unix with similar lineage as a timesharing system. It's TSO/E, and it has commands for pretty much all the functionality shown above in ISPF. Shell scripts in this environment are called CLISTS, and they have the normal sort of control flow constructs one might expect. ISPF is sort of like emacs or Midnight commander. It runs on top and provides a pseudo-graphical environment that uses 3270 full screen editing. Because mainframes have more sophisticated terminal protocols compared to Unix, the full screen UI pretty much replaced the line-based shell. It's as if emacs replaced bash entirely and gained TUI support for all system administration tasks. You can also bypass 3270 and ssh into the unix subsystem environment, and you get a pretty vanilla (and somewhat anemic) POSIX shell that behaves as you would expect. You can call out to the TSO/E shell from the POSIX shell.
Yes. Well known.
Great video Moshix. I like that this video combines lessons from your other videos in a nice digestible form. It will go in my quick reference section, for sure!
Hey moshix, thank you very much! This video explains things so nicely for total noobs :)
I had some brief experience w IBM back in early '90s (at my university), but everything was totally diff from MVS | Z/OS. That was VM/ESA and we had almost "normal" files (filename 8 ch , filetype 8 ch, filemode 2 [A-Z;0-6]) on minidiscs. We could 'q disk', 'filelist' etc. No dirs, but one could 'define' personal temp minidisc.
Datasets/records logic is still totally new to me, and thanks to you I think I'm finally beginning to understand :D
I guess the very first thing most Linux people stumble upon (even before the file system) is the difference of the 3270-terminal logic from the tty-like terminals.
Probably
@@moshixmainframechannelHi, Do you know how to connect zos with Linux, using LCS device ? I would like to access the ZOS via IP
@xxgabo yes but this is not a good place for tech support. Our discord channel is there for this purpose
OMG. That looks like pain, only pain. Why would anybody in this universe choose using this postbox like cli-gui instead of any shell with pipeing commands?
I am wondering what pre-interactive-terminal input/output devices the 3270-terminal inherited its IO-metaphor from. Were those punch carded blank forms? Cause this kind of interface is very obviously not implementable for TTY.
@eugrus block I/O is where it comes from. The line length of 80 is from the punch card yes, but the 3270 terminal paradigm comes from the concept of block I/O for efficiency
@@moshixmainframechannel thank you for taking time to answer this! What I kind of meant is: What was the first hardware implementation of the block I/O before interactive terminals? In other words, what was used to interface with OS/360 at the time when Unix was mostly interfaced with via a telex/hardcopy-terminal. Or was an interactive terminal required to connect to OS/360 from the early on?
Remote Job Entry was a precursor. It would collect a bunch of punch cards remotely and transfer them all at once (block I/o) via leased line.
Very helpful, I have a far stronger background in UNIX than the mainframe world.
Hi Moshix, I used to work on z/OS but moved onto other platforms for career reasons. I noticed everyone who ever touched MVS always remembers 3.4
Yes so true !!
Im start now work in z/OS after year of lux, z/OS is better
Plus one for worlds best editor. I was worried you were going to say ISPF
Hahahahh
Watching my as/400 and other IBM counterparts do the constant menu diving for the simplest task always drove me mad. The entire thing always looked less like they were actually doing anything directly in an OS and more like they were just using a text UI “shell” application on top of something IBM was never going to give them real access to. Even programmer options looked like so much hand holding and guarded rails driven through menus. I used to have to modify JCL for jobs back in the early 90s but was just surprised as hell to walk into a shop still using this in 2019.
Some people like it. Others don’t. It gets the job done apparently
@@moshixmainframechannel The programmers retrained onto other systems easily enough. However the “power users” who had just memorized the menu sequences (“3, 5, 1”, “1,1,6”) took forever to retrain on mouse driven menus lol.
It's interesting that z/OS requires the user to specify the actual amount of space to allocate for the file (excuse me, the "dataset") before its creation. However, I wonder if that opens up a potential security problem. Many decades ago when I was a university student I had an account on a PDP 11/70 that ran RSTS/E. That OS didn't require it but you could tell it to allocate a specified number of bytes to the file at creation time. However, this feature not only created a file ofy the size you wanted, it left it full of whatever data that was written in the free space allocated to the file! I found it fun to allocate big files and then pop it into TECO and see what non-binary text there was to see. Does z/OS suffer from this or has IBM done something about it, hopefully LONG ago?
Not an issue on z/OS
at 45:40 what you are looking for is ISPF 3.17 :)
Thanks
Very good video for start from scratch
$? holds the return-code of the last executed program in BASH. So for example echo $? shows it.
Thanks
List all the files in all the disks: you can’t to that in traditional Windows either. The reason one can in Unix is because all the directories / folders have a common root. There is no C: that van be confusing for Linux Users at the beginning.
True
"int maint" I can see you are z/VM guy. I have same problem when writing C programs, every time. :D
You unmasked me. Yes I am a VM370 guy
you can get the latest z/OS with the full suite etc.. CICS, Omegamon,RACF etc.. on the new IBM Hobbiest license pack. for 100 dollars.
Great. Link ?
Excellent video!!! Thanks!
cool video!
old school ... ? definetly
Would you consider for new projects on a green field approach any business case still to use a mainframe? I am watching you channel for days in a row and till now I can not see why I should start any project on this outdated concept. Since the introduction of Kubernetes and Micro Services like Netflix, I believe the business case got canceled for any new Mainframe
Okay
The physical mainframe machine has many advantages, like hotswappable CPUs and RAM.
Mainframes are not needed for common tasks like hosting websites, because those workloads are both easilly distributable over many small machines and non-critical systems.
hahah .. moshix .... you work odd hours.... well it is 00:43 here ...
Jet lag
find /
lists all files in linux
Thanks. Well understood
You found my user I’m PRAK855
Do you know of a system someone could log into just to learn Z/OS?
IBM has a master the mainframe program
Hi Moshix I am working a project to basically create a n Oracle VM Virtualbox interface to create linux images on a bare metal mainframe. I need some help. If
you would be interested in assisting (paid) please reach out to me. If you know of anyone else that be interested please put me in touch with them. Thanks
Sure sounds interesting. Can you Contact me by email at moshix Gmail.
@@moshixmainframechannel Will do.