PEP 3131: Supporting Non-ASCII Identifiers

[Alexander Schmolck]

Neil Hodgson <nyamatongwe+thunder at gmail.com> writes:

> Paul Rubin wrote:
>>> Plenty of programming languages already support unicode identifiers, 
>>
>> Could you name a few?  Thanks.
>
>    C#, Java, Ecmascript, Visual Basic.

(i.e. everything that isn't a legacy or niche language)

scheme (major implementations such as PLT and the upcoming standard), the most
popular common lisp implementations, haskell[1], fortress[2], perl 6 and I should
imagine (but haven't checked) all new java or .NET based languages (F#,
IronPython, JavaFX, Groovy, etc.) as well -- the same goes for XML-based
languages.

(i.e. everything that's up and coming, too)

So as Neil said, I don't think keeping python ASCII and interoperable is an
option. I don't happen to think the anti-unicode arguments that have been
advanced so far terribly convincing so far[3], but even if they were it
wouldn't matter much -- the ability of functioning as a painless glue language
has always been absolutely vital for python.

cheers

'as

Footnotes: 
[1]  <http://hackage.haskell.org/trac/haskell-prime/wiki/UnicodeInHaskellSource>

[2]  <http://research.sun.com/projects/plrg/fortress.pdf>

[3]  Although I do agree that mechanisms to avoid spoofing and similar
     problems (what normalization scheme and constraints unicode identifiers
     should be subjected to) merit careful discussion.