I'm not fond of COBOL.
Let's be honest: I hate it. "But you're a mainframe guy!" So? You assume some requisite that mainframes speak COBOL. Au contraire.
Mainframes are very chic. Even z/OS itself, the flagship mainframe operating system, has a serious Unix environment. Then there's Linux in striking contrast to the verbose old workhorse language. (Though COBOL can be compiled on Linux, mainframe or otherwise.)
But there is something good to be said for COBOL. And here it is: COBOL resists obsolescence. No ... really ... it does!.
My friend John Cook posted in G+ about the documentary "The Light Bulb Conspiracy". The trailer and even the full hour are on YouTube. You should watch it. It will be [drum roll] enlightening [rim shot].
Planned Obsolescence ... it means you buy a new HP printer every year or so rather than service or repair your current one. Don't you hate that? Time was that HP DeskJet (laser!) printers were built to last, and some of the older models have continued providing excellent service for like two decades. Go figure. Maybe ink-jet is just another part of the conspiracy.
Business and technology have a secret love-hate relationship. I blame the fear and short-sightedness of (some) businessmen. But whatever the motive, business is wary of certain technologies and not generally interested in sustainability. (Like better health, sustainable tech makes good sense but doesn't make many dollars.) Competition mitigates this ... a little. In fact, if you read between the interlaced lines of "The Light Bulb Conspiracy" it seems to require a cartel to really push for such heinous idiocy as planned obsolescence.
Heck, I make my living on Linux and virtualization. Lemme tell ya, "business" was slow to embrace either. (And FOSS is still pretty scary.)
There seems to be hope. Some businesses recognize the consumer loyalty that comes from better technology, true cost savings, and sustainability. (I work for one.)
What does this have to do with COBOL?
When I heard about "The Light Bulb Conspiracy", by conicidence, I had just completed Kurt Beyer's excellent biography of Grace Hopper. (ISBN 978-0262013109) Turns out that some of the strongest ideals that we cling to in FOSS are right in line with the admiral's way: creative thinking, fostering youth, sharing ideas, collaborative development.
One of Hopper's goals was to make computers accessible, to put programming within reach of "normal people". (my term) The CODASYL committee was formed for the creation of a common language for business programming. Businesses knew that they needed a standard.
CODASYL was the diametric opposite of the Phoebus Cartel. (the original light bulb conspirators) CODASYL member companies probably did not intend it that way. (my opinion and speculation) But again, business and technology are at odds, and in 1959 programming was a new enough art that they were kind of forced into it.
The language itself was not created by the companies or the government, but by programmers. There was one (strong!) non-programmer feature (itself pressed for by Hopper herself). It had to be English-like. This is the aspect of COBOL that I most dislike. Maybe it's the reason the language has held up under decades of maintenance. Who knows?
Other programming languages have come and gone: Pascal, Ada, BASIC, RPG, LISP, Perl, ... the list goes on an on. One language is just not sexy enough, so "they" create another. It gets popular for a while, only to be replaced itself. Those programmers using the to-be-replaced language see the folly of the change, where the languages you program become some kind of fashion statement. (And here I am, ironically, trying to get you to see mainframes as "chic", and yet tipping my hat to a very un-chic language.)
And those old bidniss programs, written in COBOL, keep trickin the machines into doin our bidding.
I don't believe Grace Hopper was consciously pursuing sustainable software. But she clearly developed and appreciation for not having to re-code things. (The Mark I was barely a computer, and wasn't even "electronic" but rather "electro-mechanical". We have it so easy by comparison.) And that's what it's all about, getting the beast to do our bidding. But once the student has been taught, we don't want to re-train it for the same task. Then the lightbulb goes on, so to speak, and we reap the benefits of automated logic. (The computer does the repetitive work.)
And the Livermore-Pleasanton light bulb still glows after 110 years. (It was made in Ohio.) It's like that bank program ... written in COBOL ... decades ago ... just keeps on working.
This doesn't mean I like COBOL.