I really think we can implement the version as it stands in that research, and add the attributes we want to them, as shvarz suggested. That is, use the two half-reaction system proposed in their article, instead of many single step reactions as shvarz has thusfar proposed. I have a feeling he felt that the more complex half-step reactions would be less tractable, but I really don't think that's the case.
We'll fudge a bit on the integration stuff for speed of course. Probably assign free nrg based on an exponential instead of linear relation so that there are a few really dense substances to serve as more permanent nrg storage Ie: body. Likewise for other attributes. A linear relationship doesn't serve well for the number of attributes we're dealing with.
If we design a metabolism DNA that controls how the DNA's requests are converted into actions, we can get really complex and fun.
I've been reading science articles all day, so forgive me if I'm particularly dense right now. :P
