Explanation of numbering & labeling system about obst HOT 1 OPEN

I am glad you like it!

1. Yes, that is correct. The slightly longer answer is that there is some friction in how BDDs are introduced, either as encodings of logical fomulae with efficient binary operations on them, or as sets of numbers with efficient set operations on them. A logical formula can be viewed as a binary function from the set variable assignments to either true or false. Then you take the assignments mapped to true, which is (once you have fixed on ordering of variables) also a set of binary numbers.

   I agree that it would be good to have that written down in an accessible place, it is on the TODO list now.

2. Yeah, labelling is hard to get intuitive. Here is a screenshot from the current development version:

   

   Currently, I do not label the nodes with the level they are on (as is commonly done), but with the set they represent, while constructing a BDD. After construction, the nodes are simply given unique labels so that they can be referred to. It is nice to see the decision labels, so I added them at the side for the next version.

The problem with labelling is mostly that you can mix and match different types of operations. For example, you could create a BDD using a list of numbers, then create one given a formula, and then calculate the union of the two. It does not really make sense to use the boolean variables used in the formula when calculating the union. Right now the nodes have informative names only during construction, where this information exists, and get assigned generic names afterwards.

But if you have any ideas how to improve the labels, I would be very interested in hearing them!

from obst.