here's a heatmap of the activity on #asciitown!
full size: http://i.imgur.com/wtuoVm9.png
brighter => more often visited
more color => more often edited
the very center is the welcome page, where you first appear.
based on data at http://ascii.town/snapshot/
work-in-progress code by @eddyb: https://gist.github.com/eddyb/56285d9403026f3c4cd35ad83f827a46
you can see a mostly-square shape of visited tiles developing around the origin
if our travel was euclidean, you would expect a circle, and if our travel was only vertical or horizontal (manhattan distance), you would expect a diamond
but since we have keys for moving diagonally at the same speed as horizontal or vertical, you would expect a square
here's the same heatmap, but twice as tall to reflect character height
full size: http://i.imgur.com/SOYHSCS.png
this makes the tiles look the right shape to our eyes (like an 80x25 terminal), but as explained previously, this distorts distances in the sense of travel time
the original image better represents how remote the horizontally-distant tiles are
if you actually did stitch together screenshots of every tile into image, it would be very large!
my 80x25 cell terminal is 640x400 pixels, and there would be 512*512 screenshots
so, a raw RGB bitmap of the entire torus would be
640*400*512*512*3 = 201326592000 bytes = 187.5 GiB
#asciitown: it's bigger than you think!
on the other hand, PNG compression would make easy work of all the empty space
for example, the last heatmap image, http://i.imgur.com/SOYHSCS.png, as a raw RGB bitmap would have size
8192*5120*3 = 125829120 bytes = 120.0 MiB
but as a PNG it's only
172645 bytes = 168.60 KiB
hooray for PNG!
each tile in ascii.town has 80x25 cells so we decided to represent each tile as a 16x5 block in the heatmap, preserving aspect ratio
this accurately shows how long it would take to get to a given tile from the origin