On Fri, May 30, 2008 at 2:14 PM, Thorsten Altenkirch <txa at cs.nott.ac.uk> wrote: > > One way would be to internally expand the cases covered by the catch-all. I know Ulf had plans to do this, for precisely this reason. Seems reasonable. One could optimise things by only expanding catch-alls if they fail to type check. -- /NAD