Hi, author here! Also positively surprised to see this on HN haha
We (well mainly Guy, if he's around) are working on an implementation, which will be made open source at some point (still rounding the edges a bit). We have very encouraging preliminary results, it does compare well wrt SDD and CUDD. There is still some ideas we would like to try, specifically for model counting.
Thanks!
The arXiv submission says the paper is submitted to SAT26, did it get accepted?
We are still waiting for the reviews. The rebuttal phase should start soon, we'll see!