[Python-3000] Python 3000 Status Update (Long!)

[Aurélien Campéas]

On Fri, Jun 22, 2007 at 07:11:14PM +1000, Nick Coghlan wrote:
> Bill Janssen wrote:
> >>> It should amount to "map(+, operands)".
> >> Or, to be pedantic, this:
> >>
> >>      reduce(lambda x, y: x.__add__(y), operands)
> > 
> > Don't you mean:
> > 
> >    reduce(lambda x, y:  x.__add__(y), operands[1:], operands[0])
> 
> This is a nice illustration of a fairly significant issue with the 
> usability of reduce: two attempts to rewrite sum() using reduce(), and 
> both of them are buggy. Neither of the solutions above can correctly 

Maybe the specification/documentation is missing some phrasing like that : 

"The function must also be able to accept no arguments." (taken from
another language spec.) ?

Better fix the documentation than blame reduce. Of course, reduce was
taken from Lisp, where lambda is not castrated and thus allows one to
write the no-argument case with more ease. Castrated lambdas limit the
usefulness of reduce *in python*, not in general.

Regards,
Auélien.

> handle an empty sequence:
> 
> .>>> operands = []
> .>>> reduce(lambda x, y: x.__add__(y), operands).
> Traceback (most recent call last):
>    File "<stdin>", line 1, in <module>
> TypeError: reduce() of empty sequence with no initial value
> .>>> reduce(lambda x, y:  x.__add__(y), operands[1:], operands[0])
> Traceback (most recent call last):
>    File "<stdin>", line 1, in <module>
> IndexError: list index out of range
> 
> Cheers,
> Nick.
> 
> -- 
> Nick Coghlan   |   ncoghlan at gmail.com   |   Brisbane, Australia
> ---------------------------------------------------------------
>              http://www.boredomandlaziness.org
> _______________________________________________
> Python-3000 mailing list
> Python-3000 at python.org
> http://mail.python.org/mailman/listinfo/python-3000
> Unsubscribe: http://mail.python.org/mailman/options/python-3000/aurelien.campeas%40logilab.fr