PEP: Defining Python Source Code Encodings

Terry Reedy tjreedy at
Wed Jul 18 03:33:11 EDT 2001

> I agree. Allowing "funny characters" in names is not something we have
> much call for yet

because the people who would really need 'funny character' names to use
Python are not currently using it!

Suppose there were a super-neat language developed in India that had about
20 Hindi-character keywords.  If it allowed the option of writing
everything else in latin chars, most of us could (and would, if motivated)
learn the keywords, without even necessarily learning the complete Hindi
alphabet and writing system.  Few of us would even think about using the
language if we had to write all program text in Hindi since we would have
great difficulty in accurately writing and recognizing random Hindi
character sequences.

The question is whether the Everybody in CP4E is to be restricted to those
who can write and read latin character names or is to be more universal.
Guido's call, ultimately.

>and it may break many tools that have ASCII-only assumptions.

Most foreign alphabets have a standard Latin transliteration that can be
produced mechanically by a computer.  Foreign-alphabet users would need
such a program in order to share their programs with us Westerners.  The
same could be used as a front end for any other ASCII-only requirement.

Terry J. Reedy

More information about the Python-list mailing list