Richard,

Try and think of this as the first step of a code breaking procedure. We are trying to develop a model which contains absolutly no information. The requirement of the procedure (to create the model, not to break the code) is that it must yield exactly the same result (the exact same model) no matter what procedure God uses to encode the original information.

Look at any model you wish (developed by some procedure you propose). Note that your procedure is to work from the encoded data. Now, apply your procedure for developing that model to a second set of data which was encoded via a different procedure. You must obtain exactly the same result you obtained the first time or your procedure for developing your model is invalid: i.e. the model is not entirly general, what you obtain depends on the underlying encoding procedure.

You see, you are the one who is presumeing the data and the encoding is defined. The issue is subtle. I still say you are trying to look down the wrong end of the telescope. You are looking at the problem of modeling a known world. Try thinking about a procedure which can be used to organize information which is totally unknown. Not only totally unknown but proccessed (encoded) by any procedure concievable.

Also, remember that there is absolutly no information outside the encoded data.

Think about it -- Dick