This optimization does *not* preserve equality between the numeric
results of the untransformed and transformed graphs when the RNGs and seeds are
equal.  The reason is that the underlying sampler methods themselves are not
implemented in Theano, so we cannot apply the requisite DimShuffle-like
operations to the intermediate samples used to generate multiple replications
and/or independent variates.

For example, sampling a normal of size (3, 2) requires a draw of size (3, 2)
from a standard normal and we can't transpose that (3, 2) array.  If we could,
then we would be able to maintain numerical equality between graphs.

These tests now assume that the C VM versions of scan Ops should be faster than
their standard Python counterparts.

This allows one to obtain the length of a fixed-length vector that has--for
example--been cast to a different datatype, squared, etc.

Fix 215

BatchedDot only supports tensor3

fixing broadcasting behavior in BatchedDot

adding test for TypeError on dimension mismatch

removing extra tests

Get rid of singleton design and most module-level variables

Provide pointer to working docs.

Update print warnings to use warnings.warn

Refactor config code

This commit also removes redundant "WARNING" and "DEPRECATION" strings and, for
the latter, sets the category (or raises) a `DeprecationWarning`.

This makes them easier to refactor and easier to invoke from elsewhere.