Been involved in scientific programming for 8 years or so... for the last 5, been trying to figure out how to improve to overall quality of the software. Most methodologies seem focused on largescale business systems, whereas scientific systems tend to be smaller. XP might just be the ticket!