Skip to content

holdenk/picomath

 
 

Repository files navigation

picomath

Picomath is a collection of small math functions inspired by John D. Cook's Stand-alone code for numerical computing. It is not compiled into a library in the usual sense (that is, something you can link to). Rather, the functions are presented as standalone snippets of code.

License

The picomath library is in the public domain. Do whatever you want with it, no strings attached. Use at your own risk.

Features

Picomath provides the following functions:

Languages

The functions are implemented in the following languages:

Usage

The picomath library is organised into subdirectories, with each language in a subdirectory. Each function (or group of interrelated functions such as Gamma and Log Gamma) are implemented in one source file. Each source file is designed with minimal dependencies, so it can be included directly into a project as is, or copy and pasted into other code.

Test Suite

The test suite consists of two parts:

  • an SConstruct file (for SCons, a command line build system) to build the test code as needed
  • a Python test.py driver at the top level, and a corresponding test.* for each language

To build the tests:

$ scons

To run the tests:

$ python test.py

The expected output is:

Checking ada... ok
Checking cpp... ok
Checking csharp... ok
Checking d... ok
Checking erlang... ok
Checking go... ok
Checking haskell... ok
Checking java... ok
Checking javascript... ok
Checking lua... ok
Checking pascal... ok
Checking perl... ok
Checking php... ok
Checking python... ok
Checking ruby... ok
Checking scheme... ok
Checking tcl... ok

The top level test.py runs each language-specific test.* in turn, communicating with the implementation using standard input and output. A simple text-based protocol is used to execute functions and return results.

About

  • John D. Cook is the original author of most of the functions themselves.
  • Greg Hewgill implemented the test framework and translated the functions to many of the included languages.

About

Stand-alone numeric code snippets

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Python 10.5%
  • Java 6.7%
  • Scheme 6.6%
  • Haskell 6.6%
  • C# 6.6%
  • C++ 6.3%
  • Other 56.7%