Re: [Numpydiscussion] NumPyDiscussion Digest, Vol 90, Issue 56
Julian, I can see the need to recognize both column and row vectors, but why not with np.matrix? I can see no need for a new operator and hope to be able to comment more fully on PEP 465 in a few days. Colin W. On 17Mar2014 7:19 PM, numpydiscussionrequest@scipy.org wrote:
Send NumPyDiscussion mailing list submissions to numpydiscussion@scipy.org
To subscribe or unsubscribe via the World Wide Web, visit http://mail.scipy.org/mailman/listinfo/numpydiscussion or, via email, send a message with subject or body 'help' to numpydiscussionrequest@scipy.org
You can reach the person managing the list at numpydiscussionowner@scipy.org
When replying, please edit your Subject line so it is more specific than "Re: Contents of NumPyDiscussion digest..."
Today's Topics:
1. Re: [help needed] associativity and precedence of '@' (Nathaniel Smith) 2. Re: GSoC project: draft of proposal (Julian Taylor) 3. Re: [help needed] associativity and precedence of '@' (Christophe Bal) 4. Re: [help needed] associativity and precedence of '@' (Alexander Belopolsky) 5. Re: [help needed] associativity and precedence of '@' (Bago) 6. Re: [help needed] associativity and precedence of '@' (Christophe Bal) 7. Re: [help needed] associativity and precedence of '@' (Christophe Bal) 8. Re: [help needed] associativity and precedence of '@' (Nathaniel Smith)

Message: 1 Date: Mon, 17 Mar 2014 22:02:33 +0000 From: Nathaniel Smith <njs@pobox.com> Subject: Re: [Numpydiscussion] [help needed] associativity and precedence of '@' To: Discussion of Numerical Python <numpydiscussion@scipy.org> MessageID: <CAPJVwB=zBazN+fiYWJeiWOL=4a9Bf2XGxJGoTT8GFTtkDUDZw@mail.gmail.com> ContentType: text/plain; charset=UTF8
On Mon, Mar 17, 2014 at 9:38 PM, Christophe Bal <projetmbc@gmail.com> wrote:
Here is the translation. ;)
Hello, and what about something like that ?
a @ b @ c > (a @ b) @ c a * b @ c > (a * b) @ c a @ b * c > a @ (b * c)
Easy to remember: the *product has priority regarding to the @product, and we just do @product from left to right. In the terminology we've been using in this thread, this is "weakleft".
An advantage of this is that most parsers do analyze from left to right.
So I really think that it is a better choice than the weakright one. We've mostly ignored this option because of assuming that if we want leftassociativity, we should go with "sameleft" instead of "weakleft". Sameleft is:
a @ b @ c > (a @ b) @ c a * b @ c > (a * b) @ c a @ b * c > (a @ b) * c
i.e., even more lefttoright than weakleft :)
Do you think weakleft is better than sameleft?
participants (1)

Colin J. Williams