We have extensively and systematically investigated different memory optimizations for the maximum a posteriori (MAP) class of decoding algorithms. It turns out that it is not possible to present one ...