Feedback

You've done some interesting background reading in coming up with this topic. It's good to see some of the earlier papers getting a fresh look, but you will need to bear in mind the shift in perspective that later thoughts about EM for software development have brought about. Bear in mind that EM is not really conceived as 'writing EDEN code' though this is necessarily an essential aspect of the overall activity. Good to pursue the connections with AI and knowledge issues - probably the latest paper on that theme (078) is also worth a look.

Your general plan looks to be a good one, but some topics, such as lookahead have never been satisfactorily addressed from the observation, dependency and agency perspective. The difficulty is that such activities involve some form of survey and integration of many observations of positions in a game, and that's difficult to express using tkeden. Beware of simply adding a large minimaxing procedure (if you can avoid this!) - or at any rate, try to consider what would ideally be best suited to the spirit of EM in this respect (even if it goes beyond what a current implementation can sustain).

It would be good to find some references to external research - e.g. on program comprehension in orthodox contexts (Durham is a particular good site for this topic). Having external points of reference will help to give more focus to your discussion of 'what was discovered through the exercises'.

Something odd about the referencing at the end (papers 071 and 076)