This predates my Mastodon accounts but it ought to be on here so here it is: my interactive blog about the domino computer.
oh no, this has done well enough on the bad website that it's escaped "pronouns in bio" twitter and reached "overpriced monkey avatar" twitter who i suppose at least will enjoy the process of repeatedly guessing random numbers until a computer tells them they've won something of no real value
for soe reason i spent my lunch hour making this out of the defunct "Matt Parker Speaks Binary Dot Com" source code and @christianp's https://three.onefouronefivenine.com/ logic:
Matt Parker Knows All Of π:
https://github.andrewt.net/matt-parker-knows-all-of-pi/
I built a thing where you can play Yahtzee knowing in advance what dice you'll get.
This week I built the Mercator Rotator.
Mercator maps famously make Africa small and Greenland big, but by changing which way around we apply the projection we can make any place we like any size we like. Spin the globe, and see what you can create.
Hey look, the hastily discontinued gay Ken doll went to MathsJam.
https://scholar.social/@bgcarlisle/101535435026541826
This is excellent. It'd make an amazing bang-bang-bang MathsJam talk but the half-hour version is also engaging and fun.
Spoilery discussion of my solution
Then I wrote the N=8 version in binary and it clicked and I felt simultaneously like a genius and an idiot.
Spoilery discussion of my solution
Then I drew out some coloured tesseracts for low-N cases and that worked well. I found you could step up to the next solution by pulling out a copy of the solution into some new dimension, giving it N new colours, and then pulling out as many permuted versions as you need to fill in the missing flips — but it didn't give me an algorithm to devise the permutations.
Spoilery discussion of my solution
I got hung up for ages on parity — because every possible arrangement you can reach by flipping one coin has the same parity, the 8x8 chessboard can be modelled as colouring a 63-dimensional cube, rather than a 64D one. Somehow I thought the loss of that one dimension would be more useful than the nice factorisability of 64. No idea why. And you can get to it that way but it's inelegant.
Finally cracked this puzzle over lunch today. Very satisfying, I think it's the first one of these Big Hard Logic Puzzles I've solved start to finish on my own.
This also means I'm no longer messing up Matt's viewer retention stats from having paused it a week ago pretty near the beginning. I'd be interested to see the drop-off in the graph at the point when they say "pause the video now and have a go at it".
@andrewt As if this weren't all confusing enough, I believe the set of points depicted here is what's technically known as collielinear.
@andrewt It's the prefix col- which sometimes occurs in latin-derived words instead of co- or com- with base words that begin with an 'l'. You see it in "collateral" and "colloquial" and even in "college" and "collegial." That said, I think the one-l version, "colinear" is widely accepted as correct these days.
Manchester MathsJam regular and occasional tamed programmer for the Nerds