You know, no-one's actually said what the problem was; it's been labelled a "bug" here and by the FT, but so far it's unclear whether the problem was an incorrectly written/functioning program, in which case, fair enough, that's a "bug", or whether the problem was that the financial model used was inappropriately optimistic in how it rated CPDOs, or even - heaven forfend! - that the beancounter(s) responsible might have been. If the latter, then that reflects not a "computer coding error", but an error of judgement that's merely been recorded by a computer. That would smack rather of the "I didn't do anything, the computer just started doing that on its own..." syndrome with which I'm sure anyone who's worked in support is familiar. All too often it seems that as soon as someone is involved in pushing a button somewhere, any problems magically become "computer problems"... WTITA? if only because I'd like to be sure that there IS one...

Butterfly flaps its wings in the Andes, causes monsoon in the Himalaya. Bug strolls across Moody's big iron, causes global financial meltdown. Also verified, 2nd Law of Thermodynamics, Kirchoff's Laws, Sod's Law.