Things have settled down somewhat and I'm still interested in talking with you about this project.
I suppose, since you've told me some of your ideas in some detail, I should tell you some of mine, so we can get a back-and-forth going.
My initial vision was essentially to create Bayesian Planning Tool. The central idea is that you input your desired goals and outcomes, and then you start working your way towards those outcomes by building a decision-tree framework, assisted by the tool.
Let's say you have a goal of winning some particular contest, for example running a marathon. This gets input explicitly as a terminal goal. Next you break this goal down into the steps needed to reach the goal, but you do it in a probablistic fashion. For example, your intermediate goals are various training and nutritional sub-goals. If one of your sub-goals is "run