I suspect this question will sink like a stone. Anyway…
When I was a kid/teen my father subscribed to SA and it described a computer game, designed to be both fun and to teach the principles of what I suppose you might call nested programs or routines.
The idea was that you had to get a homunculus through a maze. The homunculus would make certain moves based on what a homunculus in his head was doing. The homunculus in his head would make certain moves based on what the homunculus in his head was doing, and so on. Using layers of routines within routines in this way, you could get the top level homunculus to go through the maze.
It always seemed like a cute idea, and my oldest son is getting to an age where I think he might like it. I suspect the game has to be out there somewhere, but I have no idea how to describe it in order to Google it.
If the program you describe indeed existed, it was most likely described algorithmically (it’s not like Sci. Am. was mailing out disks) in the monthly recreational mathematics column. It sounds very much like something that Douglas Hofstadter would have written, so you might try to find a copy of this book, which is a collection of the columns he wrote for said column. It might also have been one of A.K. Dewdney’s columns, so you could look for his collections of columns as well; this one was the first one published, and given your recollection of the time frame, along with the fact that he took over the column in 1983, I’d be surprised if it was in one of his later collections.
No it was actually written. I recall screenshots. As you say, it certainly wasn’t something SA was giving away on free disks. I recall it as something that some academic had written and was using as a teaching tool.
It sounds like a means of teaching recursive algorithms. The top-level homunculus would imagine a sub-homunculus moving one step north, south, east and west in turn, from its current position. For each move, the sub-homunculus would imagine a sub-sub-homunculus moving N, S, E, W from the new position, and so on, recursively. Eventually, a sub-sub-sub-…-homunculus would get through the maze and pass the signal back up to the top.
Although how you’d turn that elementary programming exercise into a game, I’m not sure. Maybe try other algorithms than the exhaustive one I described?
No, not really. I recall that each homunculus was represented by a small icon, able to operate in a square space (shown on screen in plan view).
To play the game I vaguely recall that you could set up the starting location of the lower level homunculi (?) and you could set up their operating space by putting “switches” in their path. The switches controlled the movement of the homunculi one layer up and were operated by the homunculi bumping into them.
So at its most basic, maybe the top level homunculi (H1) had to go three paces forward, turn left, go four paces, turn right and then go straight to escape the maze.
So you’d set up H2 so that it was three paces from a “turn left” switch, with a “turn right” switch four paces to the right of that (say). You’d set up H3 so that it was three paces from a “turn right” switch.
On “go”, H1 goes forward three paces. Then H2 hits its “turn left” switch causing H1 to go left. At the same time H3 hits a “turn right” switch causing H2 to turn right. H1 and H2 then go forward four paces. H2 then hits its “turn right” switch, causting H1 to go right, and march out of the maze.
There were probably more different sorts of switches than I’m remembering.
I remember a game for Apple from the mid-80’s that sounds similar (if perhaps less complex), called Robotopolis (or Robotopia?). You had to complete a series of tasks to leave Robotopolis and return to your home, but you couldn’t do the tasks yourself, as the hunter bots automatically tracked down all humans. You were given three robots you could program by creating circuits with logic gates and connecting them to the robots’ sensors and motors. In some cases this involved putting robots inside other robots.
That’s very interesting Peter, but no it was a much more consciously constructed game, rather than a simple algorithm like that you link to. Plus in 1994 I was living 8000 miles from my father and his SA subscription…
I’m certain there would be. Actually, I think my father may even still have all his old SA’s. It’s just that given a 10 or more year time span and not even a clear idea of what title to look for in a content page, well, I was hoping to shortcut the process.
I found one that my Uni has a subscription to, but it only goes back to 1983, and even then it’s abstracts only. I checked all of 1983-5, but I didn’t see anything that looked like what the OP described.