Is there any value in learning mainframes?

[Nomander]

Mainframes aren't going anywhere. They still have a qualities that are highly desired even today (and are still even being put into production).

Also, cloud computing is buzzing around and may result in a transition concerning the design and function of our network offerings in the future, so "if" you decided to focus on such, you aren't going to be holding a skill set that is obsolete.

Honestly, go into what you love dealing with. This business is not a friendly one and is beyond demanding and frustrating at times. If you do not love what you do, you WILL get burned out fairly quickly in it. Besides, your success in the field will be directly proportional to your interest in it.

[rcsteiner]

Originally Posted by Holger

I would like to ask which is the best way to learn Cobol, PI/I and the these things today? Is it more with Microfocus IDE (I also heard about an Eclipse plugin but didnt found it yet) - or is it still the way direct in the mainframe - creating a dataset and then writing the program in a VI like editor? I ask because I want to learn and I could have access to a Mainframe to study but I have to pay for it. So if it is just an IDE that is used today I dont need a real mainframe - what do think?
Greeting Holger

Basic COBOL syntax is relatively easy, though older variants like COBOL 74 can have interesting debugging issues with missing periods and nested IFs that might not be too obvious.

It's the native editing and runtime ENVIRONMENTS that the COBOL programs will run in that tends to vary between platforms and require some specialized knowledge. Or lots of it.

I've written COBOL on both OS2200 (UNIVAC/Sperry) mainframes and MCP (Burroughs) mainframes, and in both cases the shop I was in used native text editors (CTS/UEDIT and CANDE respectively).

I know programmers at my old employer (well, the entity which was formerly my old employer but is now part of Delta) used to use native mainframe editors as well.

Doesn't matter that much, though, and I'd be somewhat surprised if more PC-centric development environments don't exist. I know Unisys pushes something for their mainframe lines which we don't use here. And since IBM is in love with Eclipse, maybe it's the basis for some sort of zOS IDE somewhere. I don't know IBM mainframese (aside from a decade using ISPF and related apps, anyway)...

[slackjaw]

With much respect to the big-iron people in this thread, don't do it. Go learn C#, Java, SQL, and the basics of networking. Someday when you're considering a move to wherever and you do that first search on monster.com to see the lay of the land, you'll be very thankful.

[rcsteiner]

Originally Posted by slackjaw

With much respect to the big-iron people in this thread, don't do it. Go learn C#, Java, SQL, and the basics of networking. Someday when you're considering a move to wherever and you do that first search on monster.com to see the lay of the land, you'll be very thankful.

I would learn things like Java or Perl first, of course, but there's nothing wrong with learning multiple platforms and languages, at least if you have the time and the interest.

It's not a binary problem, and many positions will require familiarity with multiple languages and an ability to learn on the fly as needed.

[slackjaw]

You're right, there is certainly nothing wrong with learning multiple platforms and languages. However there is so much to learn, so many ways to make yourself marketable, that I can't imagine mainframes being worth the effort for someone starting a career in 2010. Learn a high level language but it's mainly the concepts in various design patterns not the syntax. Learn about sql databases. Learn network security. Learn things like Model-View-Controller pattern, test-driven software development, stored procedures for either SQL Server or Oracle. So much to learn you'll never finish.

But mainframe? I don't doubt it's used or challenging or interesting, I just can't imagine recommending that to someone today.

[Deranged Pharisee]

Originally Posted by Brill

I heard a lot of stuff about where mainframes are headed. One side says their going away and it's pointless to learn it. The other side says companies do what to get off mainframes, but it is going to take a long time. Meanwhile, all the mainframe experts are starting to retire leaving companies looking for people with mainframe skills. My company is looking for someone to train on their mainframe systems. Is this something worth considering, or would I just be headed to a dead end job?

If it's for free, YES. Always good to learn something new.

But to take it seriously in the career, NO. There are people offshore working on mainframes for 15 dollars an hour

Pick up a niche skill like business objects, informatica, cognos, microstrategy, maximo, ETL......... you are hitting into a market with VERY less resources.

Mainframes, too many fish in the sea. It's like oracle and java. You can put some weight on my thoughts coz I'm a former industry insider.