One way of looking at the results of the renormalization theory is
that there is a coupling between the length scales of the problem
that extends over all the scales. I find it very interesting that while
I have tried pretty hard to decouple SlugGo from the GNU Go
internals, many things come up which make it very tempting to
couple this longest whole-board space that SlugGo cares about
with the internal and shorter length scale considerations inside
of GNU Go.
Do you have any examples of the issues that surfaced and the kind of
coupling or integration between components at different levels that you
think might be fruitful?