[Numpy-discussion] Fwd: Allowing broadcasting of code dimensions in generalized ufuncs

Marten van Kerkwijk m.h.vankerkwijk at gmail.com
Tue Jul 3 13:17:15 EDT 2018


On Tue, Jul 3, 2018 at 12:11 PM, Stephan Hoyer <shoyer at gmail.com> wrote:

> On Tue, Jul 3, 2018 at 7:13 AM Marten van Kerkwijk <
> m.h.vankerkwijk at gmail.com> wrote:
>
>> Overall, would one way to move forward be to merge the first PR (flexible
>> and frozen) and defer the broadcastable dimensions?
>>
>
> This would have my support.
>
> I have similar misgivings about broadcastable dimensions to those raised
> by Nathaniel.
>
> In particular, I wonder if there is some way to make use of this
> functionality internally in NumPy for functions like all_equal() without
> exposing it as part of the external gufunc API.
>
>
OK, so let me explicitly ask whether there are any objections to going
forward with flexible and frozen dimensions, but deferring on broadcastable
ones until more compelling use cases have been identified?

Thanks,

Marten

p.s. I adjusted the "acceptance PR" to reflect this:
https://github.com/numpy/numpy/pull/11429
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/numpy-discussion/attachments/20180703/312b4830/attachment-0001.html>


More information about the NumPy-Discussion mailing list