Looks interesting, perhaps because it seems to be a bit more down to earth than some of the other proposals where you have to derive Whitehead&Russell before you're allowed to use the + operator. But it also seems quite dead. The latest link is from 2012.