Going further, the same author of the latter tree had a post picked by staff titled “Random Snowflake Generator Based on Cellular Automaton.” Another very cool idea for this time of the year.
Anyway, what is going on is that each character takes 4 decimal digits to encode its Unicode value.
When you escape it, the characters are replaced by their unicode values expressed in decimal, so the string looks like:
At this point, it uses a regex to change it from being a sequence of 4 hexadecimal digit characters to 2 hexadecimal digit characters, so the string becomes "%77%69%74%68%..."
At this point, it escapes the string. Giving the intended script of
 A quick search says that 4E00-9FFF is CJK (Chinese, Japanese and Korean) Unified Ideographs, and that is not the only Unicode block that seems to be devoted to CJK.
Combined with the fact that that the lower ascii characters are control characters that would't be used in a script, and you end up with almost all CJK characters. If you actually wanted to be entirely CJK, it probably wouldn't be that difficult to fudge the characters around so it fits.
Another way to describe it: the spiral's uphill slope stays constant.
t*e^it wraps the spiral round a the cone |x|=t.
Yea, too much code but I'm lazy
Here's a smaller one
but it's missing the motion at the moment.
There should be a unicode "sextuple-u" character so you can express "www" in one letter (and five fewer syllables). Think of all the time and bandwidth that would save!
The nice think about dual quaternions is that they compose. So you could then connect this screw to another screw in a smooth fashion.