Good news, I've made the necessary parser breakthrough and I've now got pattern P :: pretty much anything after this point to parse as a pattern synonym type signature on a local sub-branch of my branch. So no more annoying 'pattern type' nonsense. As for the 'pretty much anything' part, I have SPJ's original proposal implemented as a proof-of-concept: pattern C :: forall b c. (; Eq b, Num b) => b -> c -> X Maybe (Maybe b) But I see that the popular opinion now seems to be moving to pattern C :: () => (Eq b, Num b) => b -> c -> X Maybe (Maybe b) which should be even easier to implement now, so I hope to finish the branch in a couple days (it probably doesn't need more than an evening's work now). Thanks go out to everyone who contributed in this little syntax bikeshedding exercise. Bye, Gergo -- .--= ULLA! =-----------------. \ http://gergo.erdi.hu \ `---= gergo@erdi.hu =-------' Define (n.) De ting you get for breaking de law.