Ubuntu has a pre-built Condor package you can download listed for karmic. That should start a standalone Condor install, or what people call a "personal Condor" install. So you can submit jobs to the queue on that machine and they'll run, in the background via Condor, on that machine.
To connect up your other machines you'll need to repeat the install, but instead of running COLLECTOR, NEGOTIATOR and SCHEDD daemons on the box you'll just need:
DAEMON_LIST = MASTER STARTD
This makes the machine just a node that runs jobs for you. And you'll need to set CONDOR_HOST macro in the config file on this machine to be the full hostname of the machine running the collector and negotiator daemons. That will put all the machines in the same pool by having them all report their availability to the same collector/negotiator.
Once you've got them in the same pool you'll need to work out how you want to move the data around between the submit node and the execute nodes: do you want Condor to transfer the data? Do you want to use a shared filesystem? The default is to let Condor try to do the file transfers, treating the entire grid as a set of loosely coupled machines with no shared filesystem between them. This is great for a quick setup, but doesn't scale as well as a shared filesystem approach. For 10 machines though, I wouldn't worry about it.
For a more detailed getting started I highly recommend reading the Condor manual. It's a bit of a book, but worth the time spent reading at least the first 3 sections.
If the manual doesn't help, the primary place for getting Condor help is their condor-users mailing list. It's activiley monitored by U. Wisc. personnel and other Condor users. I'm actually contemplating proposing a condor.stackexchange.com...