Improper Gaussians are presented this way. If both mean-times-precision and precision were zero, the
Gaussian would be printed out as Uniform.
Tom gives a good explanation of improper distributions in
this lecture somewhere between 0h34m and 0h42m.
-Y-