[Numpy-discussion] (Late) summary of PEP-225 discussion at Scipy

Perry Greenfield perry@stsci....
Wed Oct 22 16:24:20 CDT 2008


On Oct 22, 2008, at 5:15 PM, Robert Kern wrote:
>
> I would add another position (my own) to the "Arguments neutral
> towards or against the PEP":
>
> * I really only care about having just *one* extra operator, one that
> I can (ab)use for matrix multiplication. It's the only operation that
> is common enough and with one obvious implementation (I'm looking at
> you, Matlab's "\") to warrant it, IMO. Doubling the number of
> operators and special methods is not a price that I'm willing to pay
> to get it, though.
>
Note that although some of the proposals allow for a doubling of the  
number of possible operators, numpy doesn't have to use them all and  
thus doesn't need to double the number of special methods. I'd agree  
that only a few are really necessary (to matrix multiply I'd add the  
logical operators as well).

Perry



More information about the Numpy-discussion mailing list