How do you evaluate programming language features?

Python is like a friendly book; it's easy to read and write, therefore, many people I know like to use it. But, just like we use spaces and full stops to make sentences clear, Python asks us to do little things like indentation to keep the code neat. Python can also say a lot with just a little, like drawing a big picture with a few strokes. Therefore, it lets you do many cool things without writing a ton of code. But, you still need to learn a few tricks, like list comprehensions and lambda functions, to make the most of it.

Ironically, the more "flexible" and multi-paradigm a language is, it gives as many different, inconsistent and perhaps even unaesthetic (to some) ways of producing code. This is exactly the issue with PHP, a robust and high performance language and so flexible that it almost supports every single paradigm like procedural, OOP, functional, etc. And so flexible that there is no static typing and it can even silently "handle" uncaught exceptions by continuing the next statements to cover up for the "practical" programmer! Python, on the other hand, tried to be consistent in many things. Dynamic but still strongly typed, exception handling is robust. Along with mind-blowing readability, this made it an easy choice over PHP for most coders.

Many modern programming languages tend to prefer expressions over statements. This allows you to initialize structured data as a whole and avoid uninitialized and partially initialized data structures. At the same time this plays well with the pattern matching syntax. That allows for more precise control over code correctness as the compiler can tell you whenever you do not cover the full set of situations that can happen in your program. Modern languages ten to prefer using type hints after the variable name (and a colon) as this allows for the type to be left out, letting the compiler deduce anything from simple to complex types for the programmer. This allows you to keep static analysis but avoid having to write too much.

Python is not the fastest runner in the race; it's more like a steady walker. Therefore, it’s good for many tasks, but for the high-speed race, you might need to give it a boost with some tricks like using C/C++ code or a different Python implementation like PyPy.

Performance requirements can vary. Some projects may aim for maximum throughput like the good old Cobol-based mainframe software. Many projects, however aim for predictable latency. The easiest way to achieve that is to avoid garbage collection and coroutine scheduling. On the other hand, as Rust shows to the programming world, even coroutines can be made predictable by avoiding dynamic allocation. Other projects are dealing with predictable garbage collection. That means the easiest way isn't the only one available. Some languages now prefer explicit mutability and sharing as well as strict pointer aliasing rules to provide more space for the optimizer to improve the performance of your code. Cache locality is very important nowadays.

Python has a large collection of libraries and tools like a library has books on every subject. Therefore, no matter what you want to do, Python probably has a tool for it. But, with so many options, it might take some time to find the right tool and learn how to use it.

Some languages now come with an obvious choice of tooling that can help to manage the dependencies quickly. Python is traditionally one of the best examples with other dynamic languages following. However, as Rust is more of a challenger to C and C++, its tooling that is on par with Python and JavaScript is a real game changer in the world of languages compiled to binary executables. C and C++ lag behind a lot and often make your life more terrible than it would be if there was an obvious choice of tooling that would work across many platforms as well as for cross-compiation for embedded targets. If there is a future new language or even tooling for C and C++, the example of Rust would be an obvious inspiration.

Python has a great community in my country and worldwide. It is a language for many sorts of people. We would be happy to build at least a small fraction of what the Python community does but for Rust. Let's see what we can learn from them. I would also be happy to see specialized Rust communities for embedded, data science and other fields. But also in social areas in the form of Rust for women and girls as well as for complete beginners regardless of other characteristics.

Python has a big, friendly family all over the world not only in The Netherlands, ready to help each other. Therefore, if you have a question or a problem, someone likely has the answer. But, it also means you need to be part of the community, ask questions, share your knowledge, and learn from others.

Python is like a good friend – easy to talk to, versatile, and always there to help. Therefore, it’s a great choice for many people and projects. But, like any friendship, it requires time and effort to learn and grow together.