code_swarm is an experiment in organic software visualization.
See http://vis.cs.ucdavis.edu/~ogawa/codeswarm for a picture of what we want to produce.
You will need the Apache ant build tool, at least version 5 of the Java SDK from Sun, and Python version 2.4 or later. Linux users: we've been unable to get code_swarm to compile with the GNU toolchain, Sun's implementation is strongly recommended. If you're able to get it to compile with another toolchain, we'll be glad to include instructions here on how to do so.
Recent versions of OS X come with a good java compiler and Python interpreter, so all you need to do is get and install ant.
If you have DarwinPorts/MacPorts, you can simply: sudo port install apache-ant
sudo apt-get install ant sudo apt-get install sun-java6-jdk sudo update-java-alternatives -s java-6-sun
Tested with Ubuntu 8.04, probably similar in other linux distributions.
You should also ensure that Python is installed. Run
python in the terminal if you're unsure.
- download ant: http://ant.apache.org/bindownload.cgi
unpack it and add the location of its binaries to the the
PATHenvironment variable. For instance, add at the end :
download the Sun Java SDK at http://java.sun.com/javase/downloads/index.jsp
install it and add the Java binaries to the
then create a new environment variable called JAVA_HOME and set its path to
download and install Python from http://www.python.org/download/ and ensure that it too is in the
Getting the source code
A git fork of the main code_swarm repository is maintained at http://github.com/rictic/code_swarm/tree/master
This fork contains a number of patches which improve the speed of code_swarm significantly, as well as an executable, which makes running code_swarm a one-step process.
To obtain a clone of the repository, simply use:
git clone git://github.com/rictic/code_swarm.git
The main code_swarm repository however is on Google Code: http://codeswarm.googlecode.com/
To get the source, see http://code.google.com/p/codeswarm/source/checkout
With Java and ant installed, and the code_swarm source downloaded, running it on a git, svn, or hg based project is easy:
code_swarm/binto your PATH. A line like
export PATH=$PATH:/path/to/code_swarm/binin your
~/.bash_profileshould do it
Other ways of running code_swarm
While code_swarm was developed to visualize source code repositories, its input format is generic, and some have experimented with visualizing other collaborative environments, including user activity on wikis and freebase.com.
code_swarm can be invoked by pointing it at a project config file, which contains a number of options for customizing the visualization. The config file must point at a repository xml file, which contains a set of events, which each describe a file, edited by a person, at a specific time. The time is specified as the number of milliseconds since January 1st, 1970.
Example config files can be seen in
code_swarm with a given config file, use
For an experimental GUI where you can specify a svn url to visualize:
antwill build, but not run the project
ant allwill also generate Javadoc HTML documentation
ant cleanwill delete all intermediate and binary files