On Thu, Sep 7, 2017 at 6:21 AM, Chris Angelico <ros...@gmail.com> wrote: > On Thu, Sep 7, 2017 at 10:07 PM, Steve D'Aprano > <steve+pyt...@pearwood.info> wrote: >> As a practical technique, naturally using a lookup table of pre-computed >> values >> is a good solution to many problems. But you cannot say you are performing >> general computation if *all* you do is a lookup. > > Right, exactly - but my point is that the end result is often > indistinguishable. How do you calculate the sine of an angle? You > evaluate an infinite series until you're satisfied with your > accuracy... or you look up a table of sines. I learned how to use a > four-figure table in my school years, but learning how to actually > calculate them was never important.
If this were a Python program using a tree of pre-computed states and merely selecting between child nodes based on player input as opposed to actively determining the best move for the current state, would you also argue that that is not programmed? I'm reminded of a fully analog dogfight simulator I once played. All there was to it was a red book and a blue book for the two players. Each player takes one of the books and turns to page 1, which shows an illustration of a heads-up display with the enemy jet somewhere in view and a radar display and other pertinent information. Each player decides how they will react, and then the system tells each player which page to turn to. Effectively, it's a two-player choose your own adventure; or from another perspective, a state machine with two inputs. On paper, would you assert that this state machine is not programmed, yet put it too into a Python program and suddenly it is? Nobody would claim that Babbage's difference engine was capable of general computation, yet we can still talk about "programming" it in its operation. It seems to me that HTML is in a similar position. -- https://mail.python.org/mailman/listinfo/python-list
