[RFC] Statistics of non-ASCII characters in strings

[Paul Eggert]

Carlos O'Donell wrote:
> I don't know that anyone has this data.

Martin v. Löwis gathered this data for a Django application, back when the 
Python folks were changing the representation of Unicode strings.  For what it's 
worth, in that application 98.6% of strings were all ASCII, and 99.1% of 
characters were ASCII.  See:

https://www.python.org/dev/peps/pep-0393/