Add "filter_variable" mechanism in Type

Jacobian/Hessian

If one of the "then" elements is a TensorType, and the corresponding
element in the "else" branch is a compatible type, or vice versa,
convert it into TensorType, so both have the same Type, and the
output Type can be determined.

This should fix a problem reported by Ian.
[Commit amended to fix problem spotted by Fred]

It should work similarly to filter(), but on symbolic variables,
returning equivalent variables of the current Type if they are
compatible.

Fixed import crash in optimizations

CCW#37: REVIEW/TEST/DOC theano/sparse/sandbox/sp.py:RemoveO

CCW#37: modified theano.config.mode value under which assertion of application of the optimization is done.

Test using None in slices in local_subtensor_merge

Mostly I made the code more consist with the other methods in this file, and
also made `using_list` to be more literal (i.e. not to include tuples as
well)

Before we used to use the variables `using_list`, `using_tuple`, but I feel
is much more clean if we do not rely on those variables that are meant only
for re-formating the output

Following olivier's suggestion, I refractored the code that retransforms
the output in either a list or tuple

Fix ticket number (oops).

Replace refs to Trac tickets with Assembla

Update dev start guide