Pi is not the true circle constant

Here is a compelling, interesting, coherent argument for a new circle constant :

http://tauday.com/

One major gripe I have is that using his convention would require a redefinition of sin/cos/tan/etc (he hints at it in his article).

This would turn their taylor series into horrible messes.

sin/cos in particular are important as solutions to
y" = -y

I suppose at some level, you could just shrug all this off and just use the exponential function.

To be fair, I should reveal that I despise the letter tau for no obvious reason.