StoneSociety is a mechanism for deciding what. ExtremeProgramming is a mechanism for controlling how. Neither delegate authority to hierarchies of managers and middlemen. Both work in fixed-length iterations. Perhaps the two can use eachother to work on a larger scale.
How? Suppose that UserStories come from CommissionDefinitionAuction?s. Developers are brought on via CommissionRecruitmentAuction?s. This extends the concept of a UserStory; a Commission isn't just to implement a UserStory because its developers also strive to maintain and ultimately obsolete the implementation to maximize their Commission for it.
Each CommissionDefinitionAuction? would be framed in terms of FunctionalTests. This follows the qualification ethic in MarkTwain's CuriousRepublicOfGondour?. And the size of a CommissionDefinitionAuction?'s Quantum would be used to represent a UserStory's priority, and thereby the order of the CommitmentSchedule.
A StoneSociety can also control the integration and deployment of the organs that correspond to each UserStory. Here it's key to recognize, as described in the second half of StonesAndGrocers, that Auctions and Commissions can be chained as per PipesAndFilters.
An Auction directs the output of a flow of Resources; whether the resources represent physical or informatic events makes no difference. So deployments can be established by Commisioning and connecting the Commissioned organs that result from CRA'ed UserStories according to a set of Auctions. Some Auctions route, exclusively, between a number of outputs. Some route proportionally. Some sort or filter. All can be rearranged by the StoneSociety bidding processes.
What's the benefit of this? Suppose you have a large number of users; say you're implementing enterprise architecture, say a road or telecommunications network. As users' needs change, so do their bids, and so do Commissions both for stories and for developers. This way hierarchies of managers aren't needed in order to scale up. Resources are applied according to market forces and ConstructiveInterference, not political expedience and managerial authority. An enterprise run like this would work more like an organism than a mechanism. --PeterMerel
Consider expanding terms like CRA inline, and this could turn into a good single page description of how it could actually work. Make it live, write a screenplay.
Goodness, it's been almost 10 years since I worried about this StoneSociety business. If you really want to get into it I'd suggest following the cip.doc link from the GreenCheese StoneSociety page. Truthfully I think the idea still has some big gaps: StoneSocietyFlaws.