T1

Python 0.9.0 Released — Posted to alt.sources

Guido van Rossum at Google I/O, 2008
SourceAlessio Bragadini (Wikimedia Commons) · CC BY-SA 2.0 · View on Commons ↗

Metadata

Date
Decade
1990s
Tier
T1
Sources
06
Connections
01

The first public release of Python did not arrive as a download. On 20 February 1991 Guido van Rossum posted it to the Usenet newsgroup alt.sources "as 21 uuencoded parts that had to be joined together and uudecoded to form a compressed tar file," as he recalled in 2009. Anyone who wanted it had to collect all twenty-one messages, stitch them together in the right order, decode them and unpack the result. The version number was 0.9.0.

What was in those parts had been in use for about a year already, inside one research institute in Amsterdam.

Why Python Was Written: The Amoeba Project

Van Rossum worked at CWI (Centrum voor Wiskunde en Informatica) in Amsterdam — in the foreword's words, "a government-run research lab". He had joined in late 1982 as a programmer in the ABC group; when ABC was wound down, he moved to the group building Amoeba, a micro-kernel distributed operating system developed jointly with the Vrije Universiteit Amsterdam under Andrew Tanenbaum's leadership.

In his own account, the original motivation for Python was "the perceived need for a higher level language in the Amoeba project." Writing system administration utilities in C took too long, and the Bourne shell did not fit: Amoeba's primitive operations were finer-grained and different from the ones the shell was built around. The need was for a language that would "bridge the gap between C and the shell" — which, he says, was for a long time Python's main catchphrase.

He considered and rejected porting something existing. He knew Perl 3, but found it even more tied to Unix than the shell, and did not like its syntax. His tastes had been formed by Algol 60, Pascal, Algol 68 and, above all, ABC, "on which I'd spent four years of my life."

A Descendant of ABC That Fixed ABC

The Christmas-week story is van Rossum's own, from the foreword he wrote in May 1996 for the first edition of Programming Python. In December 1989 he was looking for a "hobby" project "that would keep me occupied during the week around Christmas." His office would be closed, but he had a home computer. He decided to write an interpreter for "a descendant of ABC that would appeal to Unix/C hackers."

The same foreword is unusually specific about what he kept and what he threw out.

From ABCKept or changed
Indentation for grouping statementsKept — "derives directly from ABC" and, he wrote, "one of the language's features that is dearest to my heart"
A "closed system" that was hard to extend with new primitive operationsRejected — he "decided not repeat this mistake"
Powerful data types that "turned out to be less efficient than we hoped"A target for improvement: ABC had put "too much emphasis on theoretically optimal algorithms, and not enough tuning for common cases"
All-uppercase keywords, terms like "how-to" for procedure, an integrated structured editorDropped as unsuited to experienced Unix/C programmers

His other main influence was Modula-3. From C, he says, he borrowed "only its least controversial features". And despite the Unix orientation, "the first implementation was done on a Mac."

How Python Got Its Name

Van Rossum's account of the name starts with ABC's own naming trouble: its first name, B, had to be dropped because of confusion with the older language B. For his project he "decided to under-analyze it" and picked the first thing that came to mind — Monty Python's Flying Circus. It suited what was "essentially a 'skunkworks project'", and it fitted a tradition in the Amoeba group of naming programs after TV shows. In the 1996 foreword he called Python "a working title for the project".

The snake came later and against his wishes. He resisted associating the language with snakes for years, he wrote, and gave up only when O'Reilly wanted one for the cover of its first Python book, Programming Python.

From Hobby Project to Python 0.9.0

Work began in late December 1989, and there was a working version in the first months of 1990. The first piece of code he wrote was a simple LL(1) parser generator he called pgen — still, as of his 2009 account, part of the Python source distribution and "probably the least changed of all the code."

Through 1990 the language was used by a number of people at CWI, mostly in the Amoeba group. His officemates Sjoerd Mullender and Jack Jansen were the other key developers; Jansen went on to lead the Macintosh port for many years.

The 0.9.0 posting carried a licence that was "an almost verbatim copy of the MIT license used by the X11 project at the time", with CWI's parent organisation, Stichting Mathematisch Centrum, as the legal entity. The release already contained much of what Python programmers recognise: by the account in Wikipedia's history of the language, classes with inheritance, exception handling, functions, and core data types including list, dict and str, together with a module system borrowed from Modula-3.

After alt.sources

"There was immediately a lot of feedback," van Rossum wrote, and he kept a steady stream of releases coming. He began using CVS to share the work with Mullender and Jansen, wrote a FAQ that was posted to newsgroups, started a mailing list, and in March 1993 the comp.lang.python newsgroup was created. The milestones that followed, as he listed them in his 2009 timeline:

ReleaseDate
0.9.0, posted to alt.sources20 February 1991
1.0.026 January 1994
2.016 October 2000
3.03 December 2008

He notes that the earliest dates are approximate, since he did not consistently record them.

The End of the BDFL Era

For nearly three decades, decisions about the language rested with van Rossum as "Benevolent Dictator For Life". In July 2018 he gave that role up in a short message to the python-committers list, declining to name a successor; his reasons and the full context are on his page.

The community's answer was an elected body. Under PEP 8100, the first Steering Council election ran its vote from 21 January to 4 February 2019. Of 96 eligible core developers, 69 cast ballots, and the five elected were Barry Warsaw, Brett Cannon, Carol Willing, Guido van Rossum and Alyssa (Nick) Coghlan. Van Rossum, who had just stepped down as dictator, received the most votes of any candidate — 45.

Python's later rise to the centre of data science and machine learning is a story told on the programming-languages timeline and in van Rossum's biography. The 1991 event is the smaller, odder thing: a language built to write admin tools for an experimental operating system, named after a comedy troupe, and released to the world in twenty-one pieces.

Questions this page answers

Is Python named after the snake?
No. The name comes from Monty Python's Flying Circus. Guido van Rossum, then working at CWI in the Netherlands, started the language on his home computer in December 1989 as a hobby project for the week around Christmas, when his office was closed, and chose Python as a working title. He has written that he resisted associating the language with snakes for years, until O'Reilly put one on the cover of its first Python book.
When did Python 1.0 arrive?
On 26 January 1994. The first public release, 0.9.0, had been posted to the Usenet group alt.sources on 20 February 1991.

Sources

  1. Primaryvan Rossum, G. (2009). Personal History - part 1, CWI — The History of Python blog

    The Amoeba motivation, the naming, pgen as the first code written, the 21-part alt.sources posting, and the X11-style MIT licence

    Accessed 2026-10-04

  2. PrimaryPEP 8100 — January 2019 Steering Council election

    Accessed 2026-10-04

  3. TertiaryHistory of Python — Wikipedia

    Accessed 2026-10-04

Last updated:

Share