Allowing non-ASCII identifiers

[Paul Prescod]

Brian Quinlan wrote:
>>So other programming languages make non-ASCII characters possible but
>>English-speakers do not end up paying any price. So what's the problem?
> 
> 
> If we ever need to read any Python code that uses non-ASCII characters then
> we pay a price.

We're going around in circles. In languages where non-ASCII identifiers 
are allowed you say that you've never once had a problem with them. You 
spend a lot of time in one such language and I've spent a lot of time in 
other such languages. Why would it be a problem for Python if it is not 
for them?

 > ... I don't find that argument convincing because the lack of
> widespread problems might be related to the lack of widespread usage.

Some people claim (emphatically) that they would use it and also say 
that they would use it responsibly in a way that doesn't inconvenience 
other users. Perhaps this is what happens in the C#, Java, etc. worlds.
  I'm not sure that the API necessarily needs to know anything about this
> problem. The API user just has to use the usual Python mechanisms to
> work-around the issue e.g.
> 
> someComObject.getattr('yield')(2,3)
> someRPCServer.getattr('yield')(2,3)

I think you mean getattr(someComObject('yield'))(2,3,)

It's a little verbose but I agree that it is probably a sufficient way 
to work around the problem.

  Paul Prescod