Fair enough, can I suggest you put that information near the link on the first page. I'm sure many people will think that it works at least (that is my definition of experimental) but is buggy or crashes, not that it is completely non-functional.
Can you also provide a link or information to the intended implementation specification of the BMW-256 algo? I might be able to have a crack at it when I get a some time.
Here you can find a very good explanation of the bmw hash algo.
And yes, it is from 2010. This algo exists isn´t really new.
https://www.mathematik.hu-berlin.de/~schliebn/dl/Blue-Midnight-Wish.pdflg t.