I took COBOL off my CV years ago

Story: Uni profs: Kids today could do with a bit of 'mind-crippling' COBOLTotal Replies: 29
Author Content

Mar 12, 2013
4:30 AM EDT
When I first started out I was a COBOL programmer for IBM. I took it off my CV years ago because when prospective employers see COBOL on your CV they are instantly turned off (unless you are going for a COBOL programmers job).

The issue with COBOL programming isn't the COBOL. The issue is the mainframes that they work on. I hate CICS and I hate TSO and JCL.

I don't care if they offer double my wage. I will never write another line of COBOL.

Mar 12, 2013
4:31 AM EDT
Me neither.

Mar 12, 2013
11:27 AM EDT
Funny story back when I was in undergrad. For part of my work study financial aid, I worked in the computer room at night as a mainframe operator and help desk person. Did the nightly backups and answered phone support questions for the COBOL programmers, blah blah blah...

So, you know sometimes you just go through the motions when things get really monotonous. I loaded the nightly backup tape on the big reel-to-reel tape drive and sat down at the terminal. I might have even been talking on the phone to someone. But instead of kicking off the backup job, I started the restore job - not even thinking about what I was typing. It was one of those "oh cr@p" moments as soon as the return key was pressed.

So I restored week old data. During registration. Yep.

They next few days, the student services department had to re-key in all of the student registrations for the next term by hand from the paper copies. I kept very far away from the student services department for a while after that. Fun times.

Anyway, I don't really like COBOL either but I still have my textbook here somewhere. Never say never... ;)

Mar 12, 2013
11:50 AM EDT
When I look at COBOL it's just another programming language, but heavy on typing.

Very IBM-like.

As I am told by people who balk when I say "I'm not a programmer", the language is the last thing to worry about. Work out the process, and the language becomes merely one of implementation.

So languages may be more or less useful for a particular task, but they can all do it (in theory).

I looked at Prolog once, and my mind has never been quite the same. Urk!

Mar 12, 2013
3:20 PM EDT
It's a good idea to produce a "custom" resume for each particular job or job type you apply for. If the prospective employer has nothing to do with COBOL, leave it off. But if you're applying to be a COBOL programmer, not only would you include it but you might want to feature it prominently.

Mar 12, 2013
4:49 PM EDT
@Bob -

Left to myself, I would have never looked into prolog, but in my final semester I had to take an AI course, and the implementation language was, you guessed it, prolog. It was a whole new way of looking at things, and the perfect tool for a whole class of problems which can be reduced to traversing a graph - which covers an awful lot of problems which would be extremely difficult to solve with conventional programming paradigms.

It blew my mind.


Mar 12, 2013
7:18 PM EDT
Ah, I have fond memories of a COBOL class I took at University of Michigan in the early nineties. Never used it since...

Mar 13, 2013
9:02 AM EDT
Just to hijack the thread entirely,

Back around 1982, BYTE had a thingie postulating the following:

Any number can be reduced to one through iteration. I don't remember exactly how it was worded in their example, but in pseudo code it was...

Print Number. If Number is even, divide by two. Else, multiple Number by 3 and add one. Increment Counter. If Number greater than 1, repeat. Print Counter and end.

So for 3, it would be,

3 16 8 4 2 1

So 3 requires 5 iterations.

The question was, between 2 and 500, which starting number requires the greatest number of iterations?

I lost the last implementation of this puzzle around 1990, after me and my friends had run it on every computer we'd come across.

Radio Shack Basic was the slowest, Pascal for the TRS-80 was not much faster.

BASIC, Pascal, Fortran, C, on Z80, 8088, 286, 386, Tandy, PCs, DEC cluster, and one guy who got a job at Apple actually ran it on their CRAY, but he had to increase the scope of numbers in the test to 1000 in order to get it to run in enough time for the system to measure it.

I don't think anyone wrote it in COBOL, but it was one of the things I set out to write in Prolog before Prolog broke my mind and I staggered away in pain.

Like the original Norton "si" command's output measuring the system in "number of times more powerful than an IBM XT", it was fun to watch as, over some 9 years, computing power skyrocketed.

Mar 13, 2013
10:17 AM EDT
That's the Collatz conjecture. I use it as my initial foray into a programming language, and I give myself extra credit if I can make it multi-threaded.

Interesting tidbit: 113,383 requires 247 iterations, and the sequence wanders into values beyond 2^32, necessitating something other than "natural word size" for the iterator on 32-bit CPU's.

Mar 13, 2013
10:44 AM EDT
> I use it as my initial foray into a programming language

Yep, that was the idea. Also, to see what different languages did on the same hardware.

I had Compiled BASIC on my TRS-80 Mod 3, and it actually ran very quickly.

> extra credit if I can make it multi-threaded.

Ah yes, that awkward moment when I admit, "I'm not a Programmer", preferably in the same voice as the Arnoid said, "It's not a tumor."

As an aside to this aside, I look at new programming languages from the point of view not only of "How will I learn this" but also "How would I teach this to someone who has done no programming before?"

That's one reason I really miss BASIC. Sure, it made huge assumptions, was slow, didn't do anything well, but by its simplicity it was also very easy to learn. The structure made learning both Fortran and C much easier than otherwise.

I was looking at Python recently, and realized that the only reason it made any sense is because I already knew what it was doing, such as importing modules that would be used in the program, because I'd done C and already knew "include".

I look at the tools I have today, and realize that I would be at a loss to begin teaching the principles of programming to my kids. I learned BASIC from a book, "Basic BASIC", 6 months before I had my first computer.

Is the that first step in learning programming being raised too high?

Mar 13, 2013
10:58 AM EDT
Just for kicks, a quick search yielded http://wiki.debian.org/ProgrammingLanguage

And, sadly, there is no COBOL listed.

Mar 13, 2013
12:17 PM EDT
Well crap, it had been pushed out of my mind by time.

"Why Johnny Can't Code" http://www.salon.com/2006/09/14/basic_2/singleton/

It must have been brewing around in this brain's archives somewhere, since I didn't do anything but restate it, badly.

Mar 13, 2013
12:23 PM EDT
So many COBOL bashers.

Java guys really chafe when I call Java the COBOL of the twenty-first century, but it's COBOL guys who should get pissed off.

Never mind that compiled COBOL ran on mainframes with speed and efficiency that Java can barely bring itself to dream about, the real kicker is this:

Remember Y2K? Remember all the worries about a Computing Apocalypse?

A lot of that code was 20-30 years old. A lot of that was COBOL. Still running. Hell, still runing banks and insurance companies. Just working. Day and day out for damned near forever.

Maybe Java will do as well. Maybe not.

Mar 13, 2013
1:16 PM EDT
In defense of COBOL, a proficient COBOL programmer once put it to me this way:

"Choose your symbol names well, and it reads like a novel."

Mar 13, 2013
1:21 PM EDT
> "Choose your symbol names well, and it reads like a novel."

From what I've seen of the language (very little), I'd have to say that's a perfect summation.

Mar 13, 2013
2:52 PM EDT
So, let's see...

You want to hire Old Man and the Sea, not Dark and stormy night.

Mar 13, 2013
3:12 PM EDT
Thinking about co-eds, I wouldn't mind a stark and dorm-y night. :^)

I've also seen exceptionally little COBOL, but it seemed very heavy in typing skills.


Mar 13, 2013
3:28 PM EDT
I took two semesters of COBOL for my associate's degree. I almost ended up with a COBOL job immediately afterward, but I ended up getting a PC Support job instead. I haven't used it since, though I'm sure I could get up to speed with it again without too much effort.

Mar 13, 2013
3:43 PM EDT
You must remember that COBOL was one of the very first compiled computer languages, coming down the pike in 1959. It was also for use in business, with the original idea being that professional programmers wouldn't be needed to create business applications.

As it is, those novel-like COBOL programs did make it easier to work with non-programmers in verifying that programs were doing what they were supposed to do.

Mar 13, 2013
4:04 PM EDT
After a quick search, I find it daunting how many COBOL compilers and translators there are for Linux.

Mar 13, 2013
4:50 PM EDT
> You want to hire Old Man and the Sea, not Dark and stormy night.

Having read "The Old Man and the Sea", I'm not so sure I'd agree with that. You want your program to get your information somewhere intact, after all, not get eaten along the way.

Mar 13, 2013
5:53 PM EDT
@jd --

Au contraire. Data mining is all about digesting huge volumes of data.

Mar 13, 2013
7:21 PM EDT
> Data mining is all about digesting huge volumes of data.

But you're supposed to do the digesting, Dino, not sharks. Oh, wait a minute, we're talking corporations here. What was I thinking? Your point is taken.

Mar 13, 2013
7:48 PM EDT
Being a COBOL was a decent earner back in the day.

Mar 15, 2013
8:11 PM EDT
gus3: do you still have your solutions for the Collatz conjecture? you could post them on http://rosettacode.org/

greetings, eMBee.

Mar 17, 2013
2:13 PM EDT
They aren't "solutions," as the conjecture can't be proven, only disproven (by finding an integer which never reduces to 1).

The algorithms I came up with are here:


Mar 18, 2013
9:25 PM EDT
sorry, i was thinking in rosettacode terms. in this case the task would not be "prove the Collatz conjecture" but to implement the puzzle as Bob_Robertson stated it or a variation thereof. the solutions would simply be the code.

thanks for the link.

would you mind attaching a GFDL 1.2 license: http://www.gnu.org/licenses/fdl-1.2.html to the code to make it permissible to use the examples on rosettacode?

i don't know if i'd find the time to set up the task and copy the examples myself, but i'd suggest doing that to others there unless you want to do it yourself.

greetings, eMBee.

Mar 22, 2013
9:01 AM EDT
This came through my job search filters this morning:

COBOL Programmer Personnel Services, Inc. Wichita, Kansas


Clearly the issue of COBOL has not yet been entirely resolved. :^)

Mar 22, 2013
7:31 PM EDT
They're clearly looking for the Wichita line man.

Mar 23, 2013
8:36 PM EDT
> They're clearly looking for the Wichita line man.

I thought he retired for medical reasons recently. Wikipedia seems to confirm it.

Posting in this forum is limited to members of the group: [ForumMods, SITEADMINS, MEMBERS.]

Becoming a member of LXer is easy and free. Join Us!