Do you know COBOL...opportunities await

[Rick Sykora]

I had an opportunity to work on upgrading the VA's systems that are COBOL. Have not coded in 30 years and am not sure they could pay me enough. Think I could handle the language aspect, but pretty sure I no longer have the patience anymore to deal with figuring out someone else's coding (and likely lack of documentation)!

Agree that this seems like a good case to write a code converter.

 

[digicidal]

Plus how small is the attack vector on a COBOL system? It's true security through obscurity in that context - although the datastream between systems would still be vulnerable of course. In order to carry out any kind of significant hack, you not only have to have physical access... you've got to have coded your exploit or whatever on magnetic tape (and then brought another one for the output/payload)! I'm guessing no matter how lax physical security might be (which is actually high in those areas of course) a guy in a hoodie with a roll of mag tape is definitely going to stand out.

 

[mansr]

All the really scary old hardware out there running production systems is unix or windows, and there a loads of them.

There are probably some VMS systems still running too.

 

[Omar Cumming]

Anybody for FORTRAN, which is whai I learned more than 50 years ago! Can't find my deck of punchcards though.

Cheers

 

[blueone]

Agree that this seems like a good case to write a code converter.

There are various COBOL to C-syntax converters in open source.

 

[GaryS]

What would be custom about them? Now IBM z-series mainframes, those are COBOL machines you can still buy.

I was referring to other old languages such as VMS

 

[mansr]

Anybody for FORTRAN, which is whai I learned more than 50 years ago! Can't find my deck of punchcards though.

Fortran is still used a lot in scientific and high performance computing.

 

[blueone]

Anybody for FORTRAN, which is whai I learned more than 50 years ago!

Me too. My first programming course was Structured Fortran Programming, on a CDC Cyber 170, unless my memory is faulty. (Pun intended.) The more amazing thing is that the professor who taught the course in 1978 is still teaching. Probably not that course, but still a professor.

 

[blueone]

I was referring to other old languages such as VMS

Hmmm... the only VMS I'm aware of is an old DEC operating system. IBM's top of the line programming language back then was PL/1. Now that language was not for sissies.

 

[scott wurcer]

Fortran is still used a lot in scientific and high performance computing.

It has its diehards. I remember back in 1998 or so a friend used a compiler for the Mac to do scientific image processing. I'm still convinced it was the lack of math oriented optimizers in the compilers, for instance the popular C/Pascal compiler could not be forced to use the 8 80 bit floating point registers built into every 68040 FPU.

 

[mansr]

I was referring to other old languages such as VMS

VMS is an operating system, not a language. It first ran on DEC VAX machines and was later ported to DEC Alpha and Intel Itanium. A port to x86-64 is underway. Supported programming languages include C, C++, Fortran, and COBOL.

 

[Rick Sykora]

There are various COBOL to C-syntax converters in open source.

As I said, have been away from coding for a while...

Guess they are not porting because seems like they could readily get away from COBOL. Sure they could automate testing more so today too.

 

[ZolaIII]

Hi hi hi, most of US defence grid is still dependent upon it and its not so ancient either, last visual version is from 2002. In the end that's not necessarily a bad thing in any way. Using something long forgotten is actually more secure than something popular. Pascal for instance isn't much younger and it still rules in math and scientific simulations world, it simply work's.

 

[Frank Dernie]

Anybody for FORTRAN, which is whai I learned more than 50 years ago! Can't find my deck of punchcards though.

Cheers

Me.
I learned using Fortran and wrote my first bit of code that earned me money in 1970/71. I may well still have the card deck...

 

[DWPress]

Just reading this thread has resurrected nightmares from college days trying to at least pull a C- to pass....

 

[mansr]

There are various COBOL to C-syntax converters in open source.

That's just replacing one problem with three new ones:
- Bugs in the converter.
- Bugs in the C compiler.
- Unreadable generated C code.

 

[Soniclife]

Guess they are not porting because seems like they could readily get away from COBOL. Sure they could automate testing more so today too.

That's never as easy as it sounds, and the only accurate specification they will have for many systems is the code itself. And given most of this code is doing important things, through to critical things it's usually risky, expensive and will yield little value. It's rarely ported in my experience, it tends to get thrown away for entirely new systems.

 

[blueone]

That's just replacing one problem with three new ones:
- Bugs in the converter.
- Bugs in the C compiler.
- Unreadable generated C code.

Bugs in the converter? Seems likely. Not much money in that market, I'm guessing.

Bugs in a modern C compiler? Unlikely.

Unreadable generated C code? Probably. Though have you looked at typical open source C code these days? By my standards, trash. Linux even has GOTO statements in it, and not just a few. And code comments? Oh no, we don't need no stinking comments.

 

[LTig]

Bugs in the converter? Seems likely. Not much money in that market, I'm guessing.

Bugs in a modern C compiler? Unlikely.

Unreadable generated C code? Probably. Though have you looked at typical open source C code these days? By my standards, trash. Linux even has GOTO statements in it, and not just a few. And code comments? Oh no, we don't need no stinking comments.

The emphasis is on the word generated.

 

[blueone]

The emphasis is on the word generated.

I know. I'm just being difficult. I hate poorly written and documented code, so I take whatever chance I get to bitch about it.