Full Title: If an iteration is 14 days long, and one or more of the stories assigned to the iteration/sprint end up being only partially completed when we reach the end of 14 days, what is the correct thing to do with the stories? Should I extend the ending date for the iteration? It seems that I should not close the iteration since there is still work to do in it. Whatever I do, I want to make sure that I’m not sending velocities out of whack. Can you give some advice about what is the correct approach here?
This is a fairly common issue for those who use agile processes. There is some debate on the agile message boards, but the common answer generally returns to "if it's not done, it doesn't count," which is the purist answer. This approach would mean you simply move the feature into the next iteration/sprint. You want to avoid extending iteration/sprint dates - it throws off the team's rhythm and makes future planning more difficult.
There can be situations, however, where significant progress has been made and some useful functionality has been delivered, just not all the functionality called for in the story. For these situations, it would be appropriate to Split the story. When you Split the story, the system separates the 'completed' portion of the story from the remaining work and moves the remaining work to the next iteration/sprint. You can allocate the relative portion of the story estimate to so that each iteration/sprint would get its appropriate portion of the story estimate to count in its velocity.
Splitting should not become common. Stories should be sized appropriately so that the estimate from any one story does not have a large impact on the velocity of an iteration/sprint. If your team feels the desire to split stories too often, it could be a sign that your stories are too big. If you end an iteration/sprint wanting to split multiple stories, it's often a sign that your team needs to check in with each other sooner to determine how to maximize the functionality being delivered within the iteration/sprint, which may mean abandoning one story and focusing on the ones that can be fully completed. It's much better to finish one story than to end up with two that are half done.
