I very much like Martin Fowler’s latest Bliki entry on FeatureDevotion:
The key to beating off the waterfall is to realize that, as Dan puts it, agilists value Outcomes over Features. The feature list is a valuable tool, but it’s a means not an end. What really matters is the overall outcome, which I think of as value to the customers.
An important part of this thinking is that you expect the feature list to change as the project goes on. This happens you discover new things that you can do, and re-prioritize old things. This is the essence of adaptive planning, which has always been a key indicator of agile thinking. This results a big shift in how people think about a plan. In plan-driven projects, success and failure is often worded in terms of “did things go according to the plan?” In agile projects this is a meaningless question, because plans change so often. The plan is a tool, primarily one that you use to gauge the effect of changes: “how will adding this feature affect what we do”. The plan is a tool to figure out what should fit in the FivePoundBag. If your plan’s not constantly changing, you are very unlikely to be doing adaptive planning, and hence aren’t agile.