>>11853207>I really don't know what kind of programming would necessitate writing pure python with our resorting to efficient libraries.Anything for which those libraries don't exist yet, and many things where your custom logic can't be vectorized.
In terms of things I did at university--
Computational modeling can be difficult. If you're simulating interactions between a million entities you can maybe express it as matrix multiplication or convolution or something if you're lucky and clever enough, but the most intuitive way to do it is a loop. As soon as the loop is in your code and not in numpy's code you're fucked.
I implemented certain formal logic algorithms in Python. They ended up slow. Ideally these would be implemented as a Python library in a fast language by someone else, but they didn't exist like that yet.
Nowadays I have a job where I work on an enterprise application. Most of the heavy lifting is done by Postgres, but there's heaps of intricate business logic, and there's only so much of it that can be expressed in the database. The saving grace is that while the business logic can get complicated it rarely gets computationally intensive, so it's ok to write it in Python. But sometimes it does get intensive, and then things get slow.
Your own examples are in that same category. Lightweight logic on top of already-implemented operations. But that's not universal.
I think Python is actually a reasonable choice for most programming, but if Python isn't a slow language then I don't know what is.
>It seemed like you belonged to the crowd I mentioned beforeI understand, but I did specifically say that vectorized code is fast so your reply wouldn't have applied either way.
