POCP Chapter 10: Motivation, Training, and Experience

[David Cressey]

In my view, FORTRAN was the big introduction of the 1950s.  The first compilers were for vacuum tube computers.  FORTRAN could be called "king" through the 1960s,  but by the 1970s,  newer languages like Algol or PL/1 were definitely challenging it for dominance.

When the VAX first came out,  its primary application language was FORTRAN.  That was mostly a way to make it appealing to scientists and engineers,  DEC's core clientele.
But it decidedly made the VAX look more "retro"  than it, by rights, should have looked.  

FORTRAN earned its place in history as the first widespread machine independent language.  But like all pioneering endeavors,  it suffers from a lot of flaws.  It taught later language designers a lot about how NOT to design a programming language.  

[codger]

I know it's a dirty word around here, but I think that in 1980, COBOL was still a very major player among languages used for commercial applications. It was big both in terms of number of programmers and in lines of code written and maintained.

[The Gorn]

Quote:It taught later language designers a lot about how NOT to design a programming language.

Inflexible syntax. Use of whitespace as integral to parsing the language statements. Baroque idioms, such as picture clauses (I think that's what you used for formatting printed output.) Lack of control or flow structures in the original language other than a type of for statement.

Yech. I was so happy when I graduated and I found that essentially I would never need to touch the stuff again in my career.

[The Gorn]

I worked in small shops and engineering related departments. In that environment, you get absolutely no exposure to something like COBOL.

[David Cressey]

I agree.  For most of my professional life,  I made it a point to look down my nose at COBOL.  But,  the more I look at it in retrospect,  the more radical Grace Hopper's original vision looks to me.

COBOL was definitely a major player,  especially in data-centric commercial applications.  And it still commands a certain market share.

The way I got into commercial programming without learning COBOL was by learning Datatrieve.  Datatrieve,  first on the PDP-11 and then on the VAX was one of those languages that was supposed to be for non programmers.  But it could do a relational JOIN on data stroed in two different indexed files.  Very powerful,  although deceptively simple.  

[The Gorn]

Like I said, I worked in engineering departments and I only knew of COBOL by hearsay.

Here's what I know about the culture of people coming from COBOL environments.

I worked in a medium sized DoD prime contractor in the mid 1980s. They hired a bunch of business programmers to staff an Ada language application development project. My understanding was that they, indeed, came from COBOL and similar environments.

The project probably required a lot more looking around corners, and some ability to understand hardware and unique OS constraints, than these people seemed to be able to muster. I would hear stories about how these programmers couldn't cope with APIs to do things like render graphics. (I mean, call xxx(...) to draw a box.) And programming things to handle non console communications, like to an external interface, even when it was hidden behind an API, was a stretch for them.

It seemed like they only wanted to do very simple programming... whatever that was. They seemed to break down when some kind of time dependency or multitasking aspect was involved in the programming.

They were basically only used to batch processing.

I guess that's flamebait I just posted.  

To be fair, there was probably gross mismatch of talent sets at work. They probably knew much more than the yahoos in my hardware department about databases, SQL, data normalization, and DBA than anyone we had in our group.