Dempster-Shafer and modelling beliefs about sets / hacker news