A dialect of Joy in Python.
Copyright © 2014-2019 Simon Forman
This file is part of Thun
Thun is free software: you can redistribute it and/or modify it under the
terms of the GNU General Public License as published by the Free Software
Foundation, either version 3 of the License, or (at your option) any later
Thun is distributed in the hope that it will be useful, but WITHOUT ANY
WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A
PARTICULAR PURPOSE. See the GNU General Public License for more details.
You should have received a copy of the GNU General Public License along with
Thun. If not see <http://www.gnu.org/licenses/>.
Joy is a programming language created by Manfred von Thun that is easy to
use and understand and has many other nice properties. This Python
package implements an interpreter for a dialect of Joy that attempts to
stay very close to the spirit of Joy but does not precisely match the
behaviour of the original version(s) written in C. The main difference
between Thun and the originals, other than being written in Python, is
that it works by the "Continuation-Passing Style".
As I study Joy I find that it is very aptly named. It is clear, concise,
and ameniable to advanced techniques for constructing bug-free software.
Developed by Manfred von Thun, don't know much about him, not much on the
web about Joy and von Thun (Von Thun?) See references below.
Because it has desirable properties (concise, highly factored) the
programming process changes, the ratio of designing to writing code
shifts in favor of design. The documentation becomes extensive while the
code shrinks to stable bodies of small well-factored incantations that
are highly expressive, much like mathematical papers consist of large
bodies of exposition interlaced with mathematical formula that concisely
and precisely express the meaning of the text.
The time and attention of the programmer shifts from thinking about the
language to thinking in the language, and the development process feels
more like deriving mathematical truths than like writing ad-hoc
I hope that this package is useful in the sense that it provides an
additional joy interpreter (the binary in the archive from La Trobe seems
to run just fine on my modern Linux machine!) But I also hope that you
can read and understand the Python code and play with the implementation
The best source (no pun intended) for learning about Joy is the
information made available at the website of La Trobe University (see the
references section below for the URL) which contains source code for the
original C interpreter, Joy language source code for various functions,
and a great deal of fascinating material mostly written by Von Thun on
Joy and its deeper facets as well as how to program in it and several
interesting aspects. It's quite a treasure trove.
From PyPI in the usual way, e.g.:
pip install Thun
Or if you have downloaded the source, from the top directory:
python ./setup.py install
Or you can run the package from the top directory. To start a crude REPL:
python -m joy
The docs/ folder contains Jupyter notebooks, ... TODO
§.4 Basics of Joy
Joy is stack-based. There is a main stack that holds data items:
integers, floats, strings, functions, and sequences or quotes which hold
data items themselves.
23 1.8 'a string' "another" dup [21 18 /] [1 [2 ]]
A Joy expression is just a sequence (a.k.a. "list") of items. Sequences
intended as programs are called "quoted programs". Evaluation proceeds
by iterating through the terms in the expression, putting all literals
onto the main stack and executing functions as they are encountered.
Functions receive the current stack and return the next stack.
§.4.1 Python Semantics
In general, where otherwise unspecified, the semantics of Thun are that
of the underlying Python. That means, for example, that integers are
unbounded (whatever your machine can handle), strings cannot be added to
integers but can be multiplied, Boolean True and False are effectively
identical to ints 1 and 0, empty sequences are considered False, etc.
Nothing is done about Python exceptions currently, although it would be
possible to capture the stack and expression just before the exception
and build a robust and flexible error handler. Because they are both
just datastructures, you could immediately retry them under a debugger,
or edit either or both of the stack and expression. All state is in one
or the other.
§.4.2 Literals and Simple Functions
joy? 1 2 3
. 1 2 3
1 . 2 3
1 2 . 3
1 2 3 .
1 2 3 <-top
joy? + +
1 2 3 . + +
1 5 . +
joy? 7 *
6 . 7 *
6 7 . *
The main loop is very simple as most of the action happens through what
are called "combinators": functions which accept quoted programs on the
stack and run them in various ways. These combinators factor specific
patterns that provide the effect of control-flow in other languages (such
as ifte which is like if..then..else..) Combinators receive the current
expession in addition to the stack and return the next expression. They
work by changing the pending expression the interpreter is about to
execute. The combinators could work by making recursive calls to the
interpreter and all intermediate state would be held in the call stack of
the implementation language, in this joy implementation they work instead
by changing the pending expression and intermediate state is put there.
joy? 23 [0 >] [dup --] while
-> 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
§.4.4 Definitions and More Elaborate Functions
§.4.5 Programming and Metaprogramming
§.5 This Implementation
python -m joy
|-- COPYING - license
|-- README - this file
|-- archive - info on Joy
| |-- Joy-Programming.zip
| `-- README
|-- docs - Various Examples and Demos
| |-- * - Jupyter Notebooks on Thun and supporting modules
| `-- README - Table of Contents
| |-- joy.py - main loop, REPL
| |-- library.py - Functions, Combinators, Definitions
| |-- parser.py - convert text to Joy datastructures
| `-- utils
| |-- pretty_print.py - convert Joy datastructures to text
| `-- stack.py - work with stacks
§.6 References & Further Reading
Wikipedia entry for Joy:
Homepage at La Trobe University:
Stack based - literals (as functions) - functions - combinators -
Refactoring and making new definitions - traces and comparing
performance - metaprogramming as programming, even the lowly integer
range function can be expressed in two phases: building a specialized
program and then executing it with a combinator - ?Partial evaluation?
- ?memoized dynamic dependency graphs? - algebra