In the first part of this talk I am going to illustrate ---by means of an example taken from the UPML documentation and using the formalism of channel theory--- why information actually flows in a KBS and what it means for a knowledge component to be adapted in order to make information flow more reliable.
In the second part, I shall reflect on the previous component-adaptation life cycle, representing it as a sequence of basic channel-theoretic operations/relations, and I'll show how a language of KE life cycles emerges from these thoughts. The grammar of such a language supports a hacked-up version of Dave Robertson's tool HANSEL, which traces the properties of KE life cycles or life-cycle patterns.