Gregory's Server@mhoye ha, genuflecting is right. The insistence on punchcard efficiency is unfortunately probably impossible to argue against at the time
|
10 comments
 @mhoye oh random question: do you happen to know why there was such insistence on space collating lower than alphanumerics? I keep seeing it listed as an obvious design constraint but I don't understand why  @darius spaces were the original null-terminator, all zeroes on the cards, so I suspect but do not know that hitting a space in card readers tripped long-understood shortcuts the way modern C would shortcut on "a or b", with b never getting evaluated if a returns true - the rest of this card doesn't need to be evaluated, baked deep in the hardware. @mhoye I'm grateful that the book doesn't have too many of those! But anyway, I guess I'm wondering why the collation order actually matters--do lower collated characters ultimately get evaluated by the card readers more efficiently? @darius It turns out, they keep beating that collation drum all the way through it! Sorting via mechanical processes is that important, and the idea of a compiler or even "software as a control flow operator" barely even registers as anything worth caring about until somebody has to explain _what a compiler is_ in like the second last chapter. |
@darius At one point sales of punchcards - not the card readers or collators or any other related hardware, just the cardboard cards - constituted 30% of IBM's revenue. So, yeah.