PEP-308 a "simplicity-first" alternative

[Chris Liechti]

holger krekel <pyth at devel.trillke.net> wrote in 
news:mailman.1044992059.13191.python-list at python.org:

> I am still wondering if there is a simple way to fix the current 
> ternary op "x and y or z".  Everybody knows by now that 
> this "fails" if y is a false value.  Otherwise
> it works ok and is used in today's code everywhere. 
> 
> Inspired by "do the simplest thing that can possibly work"
> i now think that 
> 
>         x and y else z
> 
> might just do it and avoid the need for a new construct.  
> It's a very minor change just for fixing the problem at hand.  
> It should be obvious what it does.  

pro:
'else' is already used in various places (even where a C programmer would 
not expect it: for, while, try..except ;-)

con:
what is 'else' exactly doing here?
at first glance it returns y if y is not False else z (oops i just wrote 
Guidos syntax ;-)
but then, it still fails if y is a function that returns False...

so how should that be implemented? i guess the parser would have to match 
the whole "and else" construct, so "or else" would generate a SyntaxError?

> if 'x' is true then 'y' is the result.  Else 'y' is the
> result.  So it doesn't matter if 'y' has a false value. 
> Yes, it looks unusal but it is a simple "non-destructive"
> change which the people who now dislike PEP308 could
> probably live with. 

i strongly dislike many of the other 'solutions' discussed. no ':<>-' for 
me please. for those who like that, there is 
http://www.muppetlabs.com/~breadbox/bf/ <wink>

i find Guidos sugestion appealing but i share your concerns about silly 'if  
y if c else z: ...' constructs... but still it's one of the best solutions 
if we need one.

> It has the positive side effect that explaining this
> ternary op variation goes hand in hand with explaining
> how and/or works in python.

hm... see my comment above. how do you explain what 'else' exactly does 
here?

i agree that it is a good opportunity to learn about short circuiting with 
'and' and 'or' as well as the non-boolean outcome of these ops.

> And it acknowledges the
> fact that we "almost" have a ternary op. Also it reads 
> left to right and the alternatives "y else z"
> read well. 

IMHO it's definetly easy to read and would fit it.

chris

-- 
Chris <cliechti at gmx.net>