Quite some time ago I presented an approach to represent BDDs more compactly. The approach could be used to represent some formulas exponentially more compact than in ordinary BDDs. However, it was not possible to prove that this representation was never more than polynomially larger than BDDs. This approach turned out to be a particular case of SDDs.

For some time I have been working to improve the approach, and I believe that this time I have an approach that works, although quite a number of aspects have not yet been proven. I will present what I have up till now.