w3juke reads audio data files from
one or more data sources (web server(s) or local files) and
streams them into a locally configured playback programs. It
will randomly select the different music tracks to be played
based off your preferences. It has a history mechanism to avoid
playing any audio tracks that have been recently played. It has
a sophisticated rating system for tracks that allows the end
user to assign ratings to the tracks, discs, or artists to
influence how often that music will be played. It also has a
highly optimized data buffering implementation to deal with
slow Web servers, broken connections, and lossy TCP
w3juke requires the GTK+ and GTK--
libraries. Playback programs are also required -- one for each
type of data stream that you want to listen to. Typically this
will be a compiled, native executable, command line program,
such as mpg123, or ogg123. Last but not least, you need to have
a source of audio data files - either from a set of webservers
or as files available through a locally mounted filesystem.
w3juke program is pretty stable. It gets
run all day long and normally it dumps core less then once a
month. Then again, this has only really been tested against a
relatively small set of web servers.
There is some reasonable configuration information available. There are currently three documents available.
This document fully describes the syntax and format of the per-user configuration file. Very useful for the end user to understand.
This document attempts to describe the format of the Index Files that w3juke user.
This document attempts to describe how to use the helper application
You can get prepackaged files from SourceForge, or if you are a bit more adventurous you can get the most recent version via CVS. Here's a brief set of instructions as to how to get the source code from CVS:
cvs -d:pserver:firstname.lastname@example.org:/cvsroot/w3juke login
cvs -d:pserver:email@example.com:/cvsroot/w3juke get w3juke
make depend make