One of the beautiful and most desirable qualities of poetry is its ability to transmit a maximum of meaning in a minimum of words.
Scientifically, this is called non-linear encoding. Most often modeled with high-order polynomial equations and coefficients. Anyone either degreed or experienced in the information sciences in our present time will know exactly what I am talking about... and they will also realize that your attempts to "crack" the code of this poem have all, so far, used linear associative models.
Mathematically, you are assuming a linear coefficient "C" that will somehow transform the given line (or lines) of a poem into the intended meaning of the author. The linear equation looks like this:
Real Meaning = C*Literal Meaning
The problem with this simplistic approach (no offense, I am speaking literally that linear encoding/decoding is indeed more simplistic) of limiting possibilities to linear associative models is that it forces you to get "creative" in how you decide what "C", the transform coefficient, must be. This, in turn, leads to you "finding" intepretations that you wish to find, and that may have little or nothing to do with the true encoded message. In general, we tend to "find" interpretive coefficients that reflect what we wish to believe about a message (or what we believe a message is saying, rather than what it really is saying).
The power of non-linear encoding is hardly ever realized by those who confine themselves to linear interpretations: Through the use of a strong (high-order) non-linear polynomial for encoding a message, you can actually accomplish two different results (both of them being purposefully intended by the encoder):
1) You can disguise the real message from all except those who know the vector key to translate the given message.
2) You can "imprint" another selected message (or sometimes multiple messages) that you wish to send to manipulate those who you do NOT want to understand the real message. This "imprint" message can be caused to appear if linear associations are used to interpret certain words used within the encoded message.
While I can understand the "excitement" of trying to interpret such codes as "predictions of the future", I think bogz has been the one to recently point at the most important aspects of "codes" we see in language today. This is the aspect of INFORMATION. And it has been discussed by several on this board for a long time. Well done, bogz.
Just my thoughts and opinions on what you are doing here. I don't care whether or not you agree, I am just bringing some information to the forefront that you may wish to consider. I, for one, believe that the use of "Pi" has a very specific purpose, even though I have not attempted any code analysis on this "poem".
RMT