Python (programming language)

• 1. Introduction

  ›

  Python is a programming language that Guido van Rossum began building in the late 1980s while working at Centrum Wiskunde and Informatica in the Netherlands. What is unusual about Python is that its name has nothing to do with snakes. Van Rossum was a fan of the British comedy group Monty Python, and he wanted the language to feel fun. That playful spirit shaped everything from its documentation to the example variable names that Python programmers still use today.

  Python grew from frustration. Van Rossum had worked on a language called ABC, and he found its approach limiting. He wanted something different: a small, clean core with a large library of ready-made tools, and an interpreter that other software could extend easily. What he built became one of the most widely used programming languages on the planet. The questions worth asking are: how did a hobbyist project from the Netherlands end up teaching millions of people to code, and what design choices made that happen?

• 2. The ABC Problem

  ›

  ABC was the language that made Python necessary. Van Rossum had worked on ABC at CWI, and he watched it represent a philosophy he found deeply wrong: build everything into the language itself, make it monolithic and self-contained. ABC was inspired by a language called SETL, and it was capable of handling certain kinds of errors and interacting with an operating system called Amoeba. But Van Rossum's frustration with ABC's opposite approach, packing all functionality into one core, is what pointed him toward Python's defining characteristic.

  Python implementation began in December 1989. Van Rossum designed Python to be extensible through modules rather than bloated at its center. That single architectural decision meant developers could add capabilities to Python without touching the language itself. It also made Python practical as a way to add programmable interfaces to software that already existed. Van Rossum released the first public version, Python 0.9.0, in 1991. The name, borrowed from Monty Python, signaled that this was not going to be another dry academic exercise.

• 3. Benevolent Dictator

  ›

  For nearly three decades, Van Rossum held an unusual title: Benevolent Dictator for Life. The Python community bestowed it on him to describe the reality of how the language was governed. He was the chief decision-maker, and he held that authority alone. Then, on the 12th of July 2018, he announced his permanent vacation from those responsibilities.

  His departure was not entirely peaceful. Van Rossum resigned specifically after a conflict over adding what is called the assignment expression operator to Python 3.8. The Python community had debated this feature intensely, and the conflict made him step back entirely. In January 2019, active Python core developers elected a five-member Steering Council to lead the project going forward. Van Rossum himself has since come out of retirement and refers to himself as BDFL-emeritus. The conflict over that one operator illustrates something real about Python's philosophy: even small additions to the language face serious scrutiny, because Python's culture treats simplicity as a value worth defending.

• 4. The Zen and Its Critics

  ›

  Tim Peters wrote a document called the Zen of Python, formalized as PEP 20, which distills the language's core philosophy into a set of aphorisms. Among them: explicit is better than implicit; simple is better than complex; readability counts; there should be one, and preferably only one, obvious way to do it. These are not rules enforced by the compiler. They are guidelines, and Python has received persistent criticism for violating them anyway.

  Alex Martelli, a Fellow at the Python Software Foundation and author of Python books, captured the cultural standard bluntly. In Pythonic culture, he wrote, calling something clever is not a compliment. The community coined the term pythonic to describe code that uses Python idioms well, reads naturally, and conforms to the language's minimalist values. However, practical pressures push against this. Python provides at least three ways to format a string literal, with no authoritative guidance on which to choose. The Zen, as the community acknowledges, describes an aspiration rather than an achieved reality. Python even uses the placeholder words spam and eggs in official documentation as a tribute to a Monty Python sketch about spam, rather than the conventional foo and bar used elsewhere in programming.

• 5. The Great Split: Python 2 and 3

  ›

  Python 2.0 arrived on the 16th of October 2000, bringing list comprehensions, cycle-detecting garbage collection, reference counting, and Unicode support. For years, Python 2 was the version the world used. Then Python 3.0 was released on the 3rd of December 2008, and it introduced a break that the community spent over a decade managing.

  Python 3 was not backward-compatible with Python 2. The end-of-life for Python 2.7 was first set for 2015, then postponed to 2020. The reason for that postponement was practical: an enormous body of existing code could not easily be ported forward. Python 2.7.18, released in 2020, was the final release in the Python 2 line. Even after official support ended, an alternative implementation called PyPy continued to support Python 2.7, adding backported security updates under the version designation 2.7.18+. Today, the Python Software Foundation actively supports versions 3.10 through 3.14. Since November 2025, Python 3.10 has been the oldest maintained branch.

• 6. How Python Reads Code

  ›

  Most programming languages use curly brackets to define where a block of code begins and ends. Python does not. It uses indentation. An increase in indentation signals the start of a new block; a decrease signals the end. The recommended indent size is four spaces. This rule, sometimes called the off-side rule, means that in Python, the visual structure of a program and its logical structure are identical.

  Python also enforces a strict boundary between expressions and statements. In languages such as Common Lisp, Scheme, or Ruby, that line is blurry. Python keeps them separate, which leads to some duplication: for-loops and list comprehensions both handle iteration, and if-blocks and conditional expressions both handle branching. The payoff is that Python code tends to be readable by people who did not write it. Python forbids certain operations by design: it has no do-while loops, which Van Rossum considered harmful. It also does not support tail call optimization or first-class continuations, and Van Rossum stated that the language never will. These refusals are deliberate. Every omission is a readability decision.

• 7. A Universe of Implementations

  ›

  CPython is the reference implementation, written in C and meeting the C11 standard since version 3.11. It compiles Python programs into intermediate bytecode, which a virtual machine then executes. CPython's energy usage for typical code is worse than C by a factor of 75.88, and its throughput is worse than C by a factor of 71.9. Memory usage runs about 2.4 times higher than C. These are real costs, and the Python ecosystem has built a range of alternatives in response.

  PyPy uses just-in-time compilation and often improves speed significantly compared to CPython, though it does not support every C-based library. Codon uses an ahead-of-time compiler and claims speedups of ten to a hundred times over CPython for suitable code. MicroPython and CircuitPython are optimized for microcontrollers, including the Lego Mindstorms EV3. RustPython, written in Rust, aims for compatibility with CPython including its C-level interface, and is already used in projects such as GrepTimeDB and Ruff. Nokia released PyS60, a Python 2 interpreter for Series 60 mobile phones, in 2005. Google launched a project called Unladen Swallow in 2009 with the goal of speeding up the Python interpreter five-fold using LLVM, though that project is now unsupported. The breadth of these efforts reflects a persistent tension: Python's design prioritizes developer clarity over machine speed, and that tradeoff invites constant reinvention.

• 8. Where Python Went

  ›

  Since 2003, Python has consistently ranked among the top ten most popular programming languages in the TIOBE Programming Community Index, which bases its rankings on searches across 24 platforms. The Python Package Index, known as PyPI, contains over 614,339 packages. Machine learning has become one of Python's most visible domains, drawing in practitioners who rely on libraries that often run C or Fortran underneath to deliver performance the Python interpreter cannot.

  Python is widely taught as an introductory programming language, which means generations of programmers wrote their first function using the def keyword and their first loop with a for statement. Languages including ECMAScript, JavaScript, Julia, Go, and Mojo have drawn direct influence from Python's design. Julia was designed to be as usable for general programming as Python. Go was designed for the speed of working in a dynamic language like Python. Python 3.15, currently in alpha, is expected to reach stable release in October 2026, and it adds the ability to lazily import modules using a new lazy keyword. The project that Van Rossum started in December 1989 is still being actively extended, one annual release at a time.