1) Install Cygwin
This is as straightforward as it gets, but don't forget to add your favorite editor : vim is not included in the default install (!).
The default directory is c:\cygwin, no reason to change it.
2) Install the Java Development Kit
You should avoid unnecessary spaces in the installation directory : c:\jdk1.6 will do nicely.
3) Install Hadoop
Download the latest release (0.21.0 at the time of writing) and extract it to d:\work (or something similar).
4) Fix your environment variables
Start cygwin and append the following lines at the end of your .bashrc file:
$ export JAVA_HOME=/cygdrive/c/jdk1.6
$ export HADOOP_INSTALL=/cygdrive/d/work/hadoop-0.21.0
$ export PATH=$PATH:$HADOOP_INSTALL/bin
5) Fix the hadoop-config script
$ vi $HADOOP_INSTALL/bin/hadoop-config.sh
Locate this section starting with "# cygwin path translation" and add the following line :
CLASSPATH=`cygpath -wp "$CLASSPATH"`
Save and exit.
6) Test your installation
$ hadoop version
That's it. Happy hadoop'ing :)