This interesting social history of computer programming, The Computer Boys Take Over by Nathan Ensmenger, is essentially a story of the struggle for power inside corporations. As I noted in my pre-view of the book, programming was initially thought of as a rather lowly support function, and it took some time for the designers of the massive early computers to realise that writing software was going to be a key function. When they did, out went the women programmers who were rather prominent in those early days (although some, such as Admiral Grace Hopper, remained influential), and in came the 'boys'.
From then on, however, the software industry was mainly characterised as being in 'crisis'. In universities, academics established computer science faculties which did what academics do – they were interested in intellectual crunchiness and publishing in learned journals (although computer science majors were established surprisingly recently, mostly from the 1980s on). However, businesses, spending from the late 1950s onwards large sums of money buying computer hardware, were horrified to find they also had to spend large sums on writing software – and what's more, there were not enough people who could do this work, and no standardisation. So upgrading the hardware meant rewriting the software, and nobody could read or adapt the software written by anybody else.
The development of standard languages was one response, Fortran and COBOL being the leading examples – the former preferred by pointy heads, the latter much more widely used in business. Training programmes were established too, and most big employers introduced aptitude tests to hire and train their own programmers (both mechanisms which helped exclude women). Yet the shortage of programmers relative to the demand for software remained such that the mystique of the 'computer boy' developed – the anti-social, unconventional geek who refused to wear the corporate uniform and tended to work at night. One psychological profile of the programmer type read: “Creativity is a major attribute of technically oriented people. Look for those who like intellectual challenge rather than interpersonal relations.”
The scene was set then for a continuing culture clash between managers in hierarchical corporations and the essential, enigmatic priesthood of the computer geeks. One reason so many businesses insisted on COBOL – regarded from the start as technically inferior – was the (incorrect) notion that anybody could read and understand a COBOL programme, so top management would be less at the mercy of the programmers. The upshot is that it is the most widely used computer language ever. For example, 90% of the world's financial transactions are processed by applications written in COBOL, and 70% of Merrill Lynch's applications (p227) – and look how well banks' managements understand what's been going on. I've also heard top bankers bemoan the fact that all their software is written in a language that nobody now studies or wants to work on because it's regarded as inferior and declining.
All in all, this is an interesting read about both the profession of programmer (up to but not including the dot com boom, which has surely changed the sociology dramatically), and about the organisation of the classic big business hierarchies. The book's unduly repetitive, with the same quotes and sentences repeated in two or three places, but that aside I'd recommend it to those interested in both the history of computing and business history.
“Fortran and COBOL being the leading examples – the former preferred by pointy heads, the latter much more widely used in business.”
I'll have you know that the very first project I ever worked on in paid employment for a business as a programmer was in Fortran and to this day I have never ever written a line of COBOL.
As an aside, the best book about programming is still the same as when I started work, Frederick Brooks “The Mythical Man-Month”.
The Brooks book is quoted in this one. I was a Fortran girl myself but have heard bankers complain about being chained to COBOL