Dear Frederico,
I greatly admire your essay. You ambitiously tackle issues that some of history's greatest scientists, from Liebniz to the founders of quantum theory, have wrestled with. Your general approach is relevant to the whole practice of science. For mathematical reasons (principally Godel's incompleteness theorem), I think that the achievement of a "perfectly closed theory" may not be possible, but I see from your comment thread that you have already considered this, and presumably the intent of your program is to achieve a theory as "closed as mathematically possible." In any case, I think that the approach you suggest should be followed as far as mathematics will allow.
Hence, you may have already thought about many of the following considerations. Please don't interpret them as criticism; rest assured that I rate your contribution very highly!
1. I am not quite sure how far one can go in the requirement that a theory be "closed." For example, general relativity invokes a four-dimensional Lorentzian manifold interpreted as "spacetime." But what is a manifold? Well, first of all, it is a set. What is a set? Well, one might use the Zermelo-Frankel axioms. However, this immediately leads to Godel-type issues. Is the question of what statements are "true" in the theory included in its "meaning?" If so, then there is immediate trouble because of Godel's incompleteness theorem.
2. To some extent, I agree with those among the quantum theorists who believe something along the lines of the statement that "quantum theory should provide us with a new worldview." However, it seems that this line of reasoning can also be dangerous, because it can lead one to dismiss as meaningless issues of "interpretation" which are actually significant after all. For instance, the Hilbert space/operator algebra version of quantum field theory and Feynman's sum-over-histories version are indeed equivalent for ordinary flat spacetime, but these versions generalize in very different ways and apply to different physical models, for instance, in quantum gravity. If a model corresponding to one version turns out to "work," while all models involving the other version fail, then it really does matter what interpretation one takes. Of course, this does not disagree with anything you are saying, since it would merely narrow the choices of "interpretation" (i.e. "worldview"), and move one towards a more "closed" theory.
3. Regarding Heisenberg's definition of a closed theory, the ghost of Godel rises again to frown on the phrase "non-contradictory fashion," and the sentence "The mathematical image of the system ensures that contradictions cannot occur in the system." Heisenberg may not have known this at the time, but mathematical formalism is no refuge from contradiction. In general, it is not possible to prove such a system noncontradictory. Leibniz's dream of a "characteristica universalis," is what Bertrand Russell and company were trying to do with their Principia Mathematica when they ran into Russell's paradox. Later Godel wrecked the whole program with his undecidability theorem.
However, regardless of whether mathematical perfection of this sort is possible, there is a vast gulf between our current physical theories and the "best that could be done" in developing a closed theory. Hence, I feel the idea and the program are well-worth pursuing.
I congratulate you for a deep and insightful contribution, and wish you the best of luck in the contest. Take care,
Ben Dribus