From: Phrogz <gavin@...>
Date: 2006-03-27T09:23:54+09:00
Subject: Re: Data Structure for n-dimension values with defaults

Cool, that seems to do it. So now, I just need to figure out how that
resolution algorithm scales in the n-dimensional case when I'm trying
to optimize for the fewest number of master values involved. And, as
Ryan Davis pointed out on IRC, what my logic is when conflicting values
have the same weight (and if I should even be able to get into such a
situation).

This is helpful, though. I was previously trying to approach the
problem from some sort of default-hash value solution. This has made
apparent that I need to be storing n-dimensional data, and only if I
find a nil value at the intersection of all axes do I go into a
separate resolution mode for searching for the best value.