One task. Fewest lines wins. Pick a fight.
Ranking whole repositories against each other never meant anything, because two repositories are not attempting the same thing. These are: every entry on a board below solves the same stated problem, so the line count is finally worth arguing about.
To enter: build the thing, push it to a public GitHub repository, then count it on the front page with that challenge selected. Your line count is whatever this counter says — same rules for everyone, documented.
Shorten a URL, then redirect it.
No entries yet.
Turn Markdown into HTML without a Markdown library.
No entries yet.
Parse JSON without your language's JSON parser.
No entries yet.
Run the Game of Life and show it running.
No entries yet.
Tic-tac-toe that a human cannot beat.
No entries yet.
Turn a folder of Markdown into a website.
No entries yet.
Ranked by lines of code — comments and blank lines are excluded, so nobody has to strip their comments to compete and nobody gains by it. Bytes break ties, which is also why putting a whole program on one line wins the line column and looks absurd in the byte column beside it. Minimum 5 lines, one entry per repository, counted at whatever commit was current when you submitted — push an improvement and submit again to replace it.
Nothing here checks that your program works. The counter counts; the repository is linked so anyone can look. That is the whole enforcement mechanism, and it is the honest description of it.
The standings rank everything counted here on things that should not matter. This page is the one that does.