10-10-2019, 10:24 PM

Sorry for the delay, but here are the (scanned) HHC 2019 programming contest entries, along with the rules page.

The basic problem was "write a program that finds all three digits numbers equal to the sums of the cubes of their digits." It's a deliberately simple problem to explain, because there have been some problems in previous conferences where I couldn't even start to formulate an approach. I was looking for clever optimizations to a straightforward problem. Oh, my, did I get those!

If I ever do this again, I'll change a few things: in particular I won't give hints like "there are four numbers that satisfy the conditions", as it allowed for optimizations based on the fact that you knew the number of answers going in, which wasn't what I was looking for. But it was in the rules, so...

I originally came up with this problem in college, when I got my TI SR-52 calculator. It's been modified over the years, especially as indirect addressing was added to HP's RPN implementation. Tne fun part of the contest was seeing all the optimizations and tricks I hadn't thought of in the last 40 years or so. I'm not bitter.

I received entries in "modern" RPN (42/DM42), "classic" RPN (15C/41C), RPL, and Prime languages. Namir Shammas submitted multiple entries and won the overall award; Roger Hill, as usual, took the best RPL award. (I should mention here that Joe Horn was dragooned into judging the RPL entries.)

An anonymous note left at my place when I was away reads "Anonymous note to judge: A smart programmer can cut search space in half by requiring first two digits to have the same even-odd parity." I have no idea if this is true but wish the submitter had included a program...

If I ever do this again, I'll use a slightly more difficult problem. In the meantime, the zip attachment has the programming contest rules and the entries received. Enjoy!

The basic problem was "write a program that finds all three digits numbers equal to the sums of the cubes of their digits." It's a deliberately simple problem to explain, because there have been some problems in previous conferences where I couldn't even start to formulate an approach. I was looking for clever optimizations to a straightforward problem. Oh, my, did I get those!

If I ever do this again, I'll change a few things: in particular I won't give hints like "there are four numbers that satisfy the conditions", as it allowed for optimizations based on the fact that you knew the number of answers going in, which wasn't what I was looking for. But it was in the rules, so...

I originally came up with this problem in college, when I got my TI SR-52 calculator. It's been modified over the years, especially as indirect addressing was added to HP's RPN implementation. Tne fun part of the contest was seeing all the optimizations and tricks I hadn't thought of in the last 40 years or so. I'm not bitter.

I received entries in "modern" RPN (42/DM42), "classic" RPN (15C/41C), RPL, and Prime languages. Namir Shammas submitted multiple entries and won the overall award; Roger Hill, as usual, took the best RPL award. (I should mention here that Joe Horn was dragooned into judging the RPL entries.)

An anonymous note left at my place when I was away reads "Anonymous note to judge: A smart programmer can cut search space in half by requiring first two digits to have the same even-odd parity." I have no idea if this is true but wish the submitter had included a program...

If I ever do this again, I'll use a slightly more difficult problem. In the meantime, the zip attachment has the programming contest rules and the entries received. Enjoy!