A Mythical Phoenix: Mainframe’s Evolution | Hacker News Reader

A Mythical Phoenix: Mainframe’s Evolution

share.org

23 pointsrbanffy14 comments

I don’t claim to be an expert here so please correct me but if the mainframe vendors want to extend the life of this technology then they should invest a bit more in helping curious explorers access them. Whether through emulation, documentation which is a bit zippier (i.e., more Whys Guide to Ruby and less 1978 phone book, etc...). The closed nature of information on mainframes means that it’s hard for people to catch the bug of interest in them by fooling around first.

Mainframe emulator: http://www.hercules-390.org/

Yeah, it's kind of hard to get into the mainframe business. Kind of like getting into piloting a passenger jet. You can't pilot a passenger jet without a few hundred hours of flight time on a jet aircraft.

You can't just buy a used one and fly around on your own. So you hope someone who owns a jet will lug you around until you become competent and certified.

I know, I was just reading the part of the article that talks about how it's so hard to find people. I would love nothing better than to work on a mainframe all day long. PM me if you're hiring someone who's never touched a mainframe ;) I won't hold my breath though!

A friend of mine started playing around with Hercules and got a job with an IBM partner company. He just took a new job now at 2.5x the money I'm making, and I make good money :)

The companies who do mainframe gigs know the market is small and they would probably accept new talent that has no real world experience. All that's required is some experience (emulator counts), and a job history that says you're not an idiot and can get the job (any job) done. My friend was just a regular linux sysadmin before he got into mainframe gigs.

Also a willingness to debug COBOL programs written in the 80s to deal with stuff like Unicode on imported records.

Is debugging 90s SBCS Win32 applications dealing with importing Unicode records applicable prior experience here?

There are line-by-line debuggers for COBOL, with breakpoints, variable inspection, etc. It's not very different from debugging anything else.

It's not quite that bleak.

https://blog.adafruit.com/2016/04/05/ibm-z890-when-an-18-yea...

I have a z114 and a z800 and a P/390. Not any harder than for instance owning professional wood working gear.

And you have media for the OS you can install on them? And setup and configuration to a useable system is as simple as the guided wizards other OSs provide?

Yup. It's not like installing Ubuntu on a laptop but I don't think that's to be expected. IBM does have zPDT which is a lot easier.

IBM has a program called Master the Mainframe [0] for students. It is an opportunity to gain experience with mainframe programming on their systems. There's a competition for students, but anyone can register to for it apparently.

[0] https://www.ibm.com/it-infrastructure/z/education/master-the...

Whilst it may be true that it can run cobol, one must have to aware the assembler part of the story. Still remember the enjoyment of from 24 to 31 bit. And apar and s/370 assembler language patch.

It is not all colour.

Btw the claim you cannot run old msdos is over stretching. Still running turbo pascal on dos.

Waxing nostalgic: I loved working on mainframes back in the day. My job was a little unusual for a mainframe programmer, so maybe that's why: my team was customizing the VM/370 hypervisor and its usual guest OS ("CMS") for our employer's needs. It was mostly 370 assembly, for the OS changes, and the occasional script (or "Exec", as it was called in CMS). I wasn't doing much app development and never saw anything that looked even remotely like JCL.

I loved it, and I would never go back to it.

Reply on news.ycombinator.com