Phillipe Kruchten gave a talk today titled "Scaling Down Large Projects to Meet the Agile �Sweet Spot�". He described a way to marry the needs of large project development processes with agile development and a small-team oriented approach. In general, he suggested breaking down large projects into small teams and treat each team as an more-or-less independent, agile team with locally iterative processes and targets and a number of "management" teams that focus primarily on planning, integration and high-level customer interation and requirement development.
As far as models for integrating learning and feedback into HCI processes, I think that this methodology has a great deal of promise. Brought into software development in the context of "End-User Software Engineering it seems to me to have great deal of potential for general notification strategies and developing adaptive user interaction modes. Definitely a lot of promise.
There has often been confusion in the XML community about the relationships between names, namespaces and the binding of values. The confusion arises from a common understanding that a name is used as an identifier for some thing, in essence it is an association between an identifier and an object. Unfortunately, this is simply not what an XML name is. It is only the identifier. As a datatype, it is a constrained-format, immutable character string that is usable as an identifier only. Since we prefer short identifiers, namespaces were introduced as a way of providing a context for these identifiers (i.e. the same name can have different significance in different contexts). XML treats the binding of names to values as a completely separate phenomenon.