Modeling is supposed to be one of the most important activities in any software development process. At least this is the general understanding within the software engineering research community. However, in the day-to-day practice, modeling is usually regarded as, basically, a waste of time.
Regular readers will know that the gap between academic theory and industrial practice in this area is one of my pet obsessions (for the record, I'm on the side of working programmers). Please check out the site and join the discussion.