案例库 · 工程与运营 · 技术决策 · 2006–2020
这条还没译成中文，下面是英文原文。
NumPy made the array the basic unit of scientific computation.
NumPy gave Python a fast, expressive array object, so vectorised data operations became readable, compact and fast.
NumPy · NumFOCUS
那一手
Python was easy to write but its loops were too slow for large scientific data, forcing a choice between readable code and tolerable performance.
NumPy introduced a multidimensional array object with an efficient in-memory layout and vectorised operators, letting operations on whole arrays run fast while the syntax stayed close to mathematics.
It became the foundation of the scientific Python ecosystem, with libraries and tools built on top of it, and it was used in major efforts from gravitational-wave discovery to black-hole imaging.
为什么管用
- Vectorised operations on arrays replace slow Python loops with C-speed calls
- Array syntax stays close to the mathematics, so code is readable and compact
- A single common array object becomes the interchange format between libraries
- Being the base of a large ecosystem makes it a de facto standard
可以搬走什么
A good abstraction moves the user's effort from the mechanical details to the problem: give people a primitive that is both expressive and fast and they will build the world on it.
后来呢
NumPy became core scientific infrastructure, the base of the scientific Python ecosystem, and the interchange layer for array-based computation. It also became a lesson in sustaining an open-source project, moving from informal development to formal governance and sponsored funding to address maintenance and growth.
资料来源
发现哪里写错了？告诉我们。