Proto-PEP: Overloadable Boolean Operators

[Tim Hochberg]

Greg Ewing wrote:
> Andrew Durdin wrote:
> 
>> I'm not at a machine with the patch installed on it at the moment, but
>> I just began to whether this patch would have an effect on expressions
>> like (a < b < c)
> 
> 
> No, it won't have any effect on those. I don't think
> it should, either -- it should only apply when you
> explicitly write an 'and'.

[I should probably download the patch and try this before commenting, 
but no time, so: open mouth, insert foot].

Why? You specifically mention Numeric/Numarray as one motivating factor 
for this patch, and as a long time Numeric/Numarray user I'd rather have 
(a < b < c) work than (a and b).

My biggest concern is that if (a < b < c) is not changed in conjunction 
with (a and b), the behaviour of the former becomes difficult to 
explain. Currently one can explain it by saying, "a < b < c is 
interpreted as (a < b) and (b < c), and 'and' doesn't work on numeric 
arrays". However, if 'and' starts to work sensible for numarrays, but (a 
< b < c) retains the old behaviour, the behaviour becomes both hard to 
explain and hard to understand.

I'll try to download the patch and comment more completely later.

-tim