Wand's simple objects
I finished Wand's 1987 paper 'Complete type inference for Simple Objects'. I am a little puzzled about the corrigendum, which more or less says that Wand's simple objects actually can have multiple types at the same time. This is not taken into account at all in the original paper. I haven't seen anywhere a type system that produces multiple well-typings for a single expression. I don't know whether this corrigendum raises an interesting theoretical problem that I can forget about for my (practical) work, or whether I actually have to implement yet another algorithm.
Since I didn't find anything that looked like the promised revised version of Wand's paper (and since 1987, he would have had enough time to write one if it was really important), I ignored for now this problem and started the implementation of the original algorithm later that day. It is not ready yet, but I know more or less what to do.