Cashing in on COBOL

[SeanT]

Practically everything you can do on another system, server wise, you can do on a mainframe.

Quick, someone tell Google.

[Sausage Surprise]

I assume these are good money? Is the rate going up? Also, I assume these are all in London?

Mainly in the North East has been my experience. However, the rates were rubbish. I worked as a COBOL programmer on those systems back in 1986 and probably 90% of those I worked with have now retired. Just waiting for them to come back with a decent rate.

[malvolio]

Quick, someone tell Google.

Operationally there is no difference between a server (even a virtualised one on a large blade host) and a proper SMMP-capable mainframe. In many ways, modern client/server solutions are simply re-using the old mainframe/dumb terminal solution with better graphics.

The advantage of the modern approach is you can add CPU power a hell of a lot more cheaply.

[coolhandluke]

Well, none of this is convincing me to make the leap! I just hope all the COBOL based companies do their migrations before all the old school COBOL devs retire.

Looks like I'm going to have to stick with all the cutting edge stuff.

I believe that CGI have simply off shored all their COBOL development.

[darmstadt]

Operationally there is no difference between a server (even a virtualised one on a large blade host) and a proper SMMP-capable mainframe. In many ways, modern client/server solutions are simply re-using the old mainframe/dumb terminal solution with better graphics.

The advantage of the modern approach is you can add CPU power a hell of a lot more cheaply.

Although I do recall GIF and JPG on a dumb terminal back in the early nineties and now I can do graphics on a mainframe with Linux (or even Windows) running there (admittedly the dumb terminal is not a requirement anymore.)

The mainframe I was recently using had as a base, 141 processors and 10TB of memory but you have to remember that mainframes have been using virtualisation for a long, long time which means I had a heck of a lot of virtual processors.

[SeanT]

Operationally there is no difference between a server (even a virtualised one on a large blade host) and a proper SMMP-capable mainframe.

Operationally there's a whole world of difference between good quality cloud native distributed computing, and putting all your eggs in one basket.

[malvolio]

Operationally there's a whole world of difference between good quality cloud native distributed computing, and putting all your eggs in one basket.

Operationally I would disagree. I ran a clustered mainframe that delivered a service with no downtime for four years, using early 90s technology. Cloud simply makes it much easier to achieve.


Then again, "Cloud" is a myth. It's simply someone else's computer network. You might as well use your own and after 3-4 years you're ahead on the ROI curve.

[SeanT]

Operationally I would disagree. I ran a clustered mainframe that delivered a service with no downtime for four years, using early 90s technology. Cloud simply makes it much easier to achieve.

So you put all your eggs in a couple of baskets. Hope they were on different sites.

Then again, "Cloud" is a myth. It's simply someone else's computer network. You might as well use your own and after 3-4 years you're ahead on the ROI curve.

I know this, try telling the stupid accountants. Where did I say public cloud?

[yMyjgT]

PRO IV, now there’s one I’m not sorry to see the back of...

There's a couple of companies you should avoid working for then, still relying upon it.

[RedKiteWien]

Maybe by the time I am 70 I will be the only Mainframe Dinosaur left and will be earning 2 grand a day fixing COBOL progs and dodgy JCL ! only 16 years , then one day Rodney we will be millionaires