Don't Shoot The Messenger

Reading a short piece in the latest US edition of Wired that dropped in the mail today, about COBOL [the ancient programming language that still occupies the bulk of extant working computer code on planet earth], I was struck by the tone of the argument - echoing a not particularly recent sentiment, either - that COBOL is essentially the spawn of the devil. The analogy used is with asbestos: at once ubiquitous and as dangerous as hell and difficult to eradicate, to boot. Well, yes, this is indeed the case: COBOL is pretty much still at the heart of mainframe business, banking and government administration computer systems to this day, and yes, the numbers of competent individuals left to deal with its many-fold exigencies [blog posts passim] are few and far between. However, whilst the piece outlines the historical and pragmatic expediencies that led to its widespread adoption by governments and agencies worldwide in the first place, its demonising of the COBOL language itself [and in fairness it was most definitely not without truly manifest flaws] the essay misses the fundamental point that the language itself [as is any computer language, for that matter] is by definition agnostic of the uses to which is put. The fatal flaw, as always is that twisted combination of inadequate planning, simple economics applied simply, and a wilful lack of understanding on the part of the adoptees. In short, the suits wanted to impress the suits that sat above them in the pecking order, and who wanted to demonstrate that they were fully on the case, whilst simply delegating madly downwards and sideways, all the while demanding unrealistic goals be met by the engineers tasked with actualising their unresolved thoughts. Result? As always, compromise. And as always, when you adopt a compromised system, it stays compromised and never gets properly fixed until it's too late. Problem? Human, not software...