> Well, maybe it was, in the days of ZX Spectrum, when the machine was in > the exactly same state every time you turn it on - it read everything it > needs from a tiny EPROM, which never changed, and its address space was > 64 k, which was aboute the size a(n average) human could comprehend. actually the Spectrum had 48K and the ZX81 had only 1k (expandable to 16 ) (s) > Ever since we crossed the, let's say, megabyte border, we can't exactly > track what's going on inside. They might have introduced little green > gremlins inside the disks, and we'd never know. For every nasty thing > they do, we wonder if it's a bug or feature. I don't really think so , while I agree that it is (almost - just in case) not posible to knwo everything about what happens in your computer but the fact that standards for building software and hardware are not as high as they should be doesn't mean outcomes are not predictable - every bug has it's inner logic (g) there are filelds in CS which are on the boundaries of "exactnese" like fuzzy logic etc.but they have rules as welll Arnon