Anyway, Scrum, XP, and all the rest I've come to understand as each a record of a reaction against some bad condition, with transitional practices to get a team away form that condition to a better one. Fo example, as Michael Feathers says in this post, regular, fixed-length Iterations require (and enable) a certain kind of discipline, force a certain set of tradeoffs. And maybe doing that can help a team a lot. And maybe not, as the case may be.
That doesn't mean that, once the use of that practice has taken the team away from the bad condition then no further change to practices can be beneficial. In the specific case of regular fixed-length iterations I mainly see the application being to teams moving from a condition where no-one ever has any idea at all when anything is going to be delivered to a condition where everyone always knows a date when something is going to be delivered. In many settings that would be considered a major improvement by those paying the team. And consistently working that way is a great way for a team to gain the trust of the business.
Once that trust is established, and once the conditions are in place for frequent delivery, what need for the fixed-length iterations? I suspect that what worries a lot of people who've seen the transition from chaos to iterations is that they can't imagine a world without iterations which is not also a (return to) chaos. Michael suggests an experiment:
Suppose that you had an iteration of one week, followed by an iteration of 2 days, followed by an iteration of 1 day, followed by an iteration of one-half a day, and so on. If you still had your sanity at the end of this process, would you have learned anything? I haven’t tried it with a team yet, but here’s the thing that I hope would come across: if you apply enough ingenuity and you’ve acquired enough skill, you can deliver business value in shorter times than you can currently imagine.That would be cool. Of course, the Kanbanista's seem to suggest going straight to that world in one step. And maybe that can work in a certain setting, and maybe not. The idea scares me, whe I look at most of the teams I help.
The aspect of this sort of thing that really interests me, though, is this: if what the Certified Scrum Masters say about software development living on the “edge of chaos” is right and if what the Cynefin people say about that being exactly the place where “emergent practice” lives then by their own argument, we would expect Scrum to consist of mostly transitional practices.
Fixed-length iterations (excuse me, “sprints”) seem like a good candidate to be one.