# Proposing the numpy tag

Have proposed the numpy tag.

Currently searching on numpy yields 423 results and numpy [python] 322 so there is some interest. Not that all of these would be tagged, some IMO would benefit

Replace matrix @ vector list comprehensions with something more efficient

Has been my experience asking Blender related numpy questions on SO SO often generate little or no interest.

Indexing back to the one from the many

in relation to selecting verts using UVs selected

I knew what I wanted to achieve Blender wise, coming up with the numpy jargon to ask a good question on SO not so much. IMO Blender related numpy questions will help bridge that divide for all, and in the end help us produce faster code when dealing with large arrays (like a mesh)

To reiterate: Tagged questions would be about solving a blender specific question using numpy. The python bundled with Blender includes numpy. (It hasn't always) On the con side I understand there are many other python modules shipped with Blender & why add just one as a tag.

Anyway will leave it at this. Interested to hear your thoughts.

• Well I'm not a coder, I can't judge how useful it would be by myself, but I trust your judgement. There does seem to be a demand, and if it helps find those questions more easily, then I'm up for it. Mar 8 '21 at 23:26
• I'm for it just because seeing the word numpy makes me laugh. Mar 8 '21 at 23:38
• Does the mathematics tag not already serve this purpose? It seems to me the SE philosophy is that if the question is about numpy then it should belong on SO, if it is simply about something that may use numpy in Blender, the existing mathematics tag should cover it, right? Mar 9 '21 at 16:33
• Does python + mathematics imply numpy ? ( leaving out the moot point re what maths questions --> Maths SE :-/) TBH I'm not 100% sold on the idea of the tag, was talking myself out of it while writing the question. Arrived here after writing a question pertaining to blender, requiring a numpy solution. Followed SE philosophy and asked on SO. My conclusion from that adventure: it is better asked & more helpful here._ie_ , when I find myself, or see others asking "How to do Bar in Blender using Foo" often enough, perhaps a Foo tag is worth considering . Mar 9 '21 at 18:09
• @batFINGER to your first question: no, not necessarily. But I guess I am struggling to see a scenario where numpy specifically is necessary to be tagged and not mathematics in general. Mar 9 '21 at 21:21
• I don't think, for the tag, the distinction would lie so much in the maths that numpy does, as in the idiom it uses to do it. Is it tricky to find the line between BSE 'Blender Numpy', and SO 'Numpy Numpy?' BTW, my introduction for beginners has been through Rich Colburn's generous Numpy for Blender series of videos.. covers some of that idiom, but doesn't dig into the kind of optimisation question that you've been posing recently. Mar 10 '21 at 8:15
• Thankyou & agree it's tricky (and sticky) attempting to find the demarcation. Will check them out. Recall the name from my BA days. Mar 10 '21 at 8:20
• @batFINGER I think you're probably ahead of them.. but a really nice intro for folks like me. Mar 10 '21 at 8:23
• FWIW I'm in favor. Dealing with Python (notably slow but accessible language) and a software with many lists that can contain many vectorized elements, having easier access to questions / answers about a well maintained and optimized library will definitely help aspiring coders in the long run. Mar 12 '21 at 12:18
• I'm not a coder, and can't too well judge the usefulness but I don't really see a reason not to have it. The SE description of a tag is... A tag is a keyword or label that categorizes your question with other, similar questions. Using the right tags makes it easier for others to find and answer your question. anyway.. So it would make sense to have the tag if it is going to help other users find the right questions and answers.
– Timaroberts Mod
Mar 12 '21 at 19:39

Numpy is incredibly powerful and can solve a lot of calculation conundrums and provide capabilities that may be useful in surprising ways, but it takes getting used to. Having a numpy tag allows readers to quickly find many examples of numpy-based solutions and read through them, and this makes it a lot easer to modify them and explore new uses.