What begins to make less sense is treating dialogue as if it were a practice. In other words creating scenarios like: dialogue > purpose > understanding or, dialogue > understanding. This, of course, is every teacher’s fantasy!
It’s also the fantasy of requirements gathering during software development. I agree completely that practice needs to occur early in that chain of progress. This is the reason that I’ve been on this big prototyping quest for the last few years. Until you start doing, you don’t start understanding. In that post Chris says that talking during shared practice acts as “a kind of carrier wave of understanding”. The goal in prototyping is to move as much of the dialog from speech into the model itself, so that communication is increasingly occurring through the model instead of around it. Spoken dialog is a good way to start out, but too imprecise to facilitate understanding. I hope that’s what Chris is saying at least, cause I agree completely :-)