Parsimony branch lengths

[willfischer]

I don't think it is currently possible to compute parsimony branch lengths on user trees (i.e. using --evaluate and --tree with an input tree). Could this be made possible?

The two pitfalls I see: 1) "parsimony" cannot be specified as a model, and 2) parsimony trees may have multifurcations. Possibly (1) can be simulated by using just the JC model, and (2) by randomly decomposing each multifurcation to a series of bifurcations, but -- is it possible that this is a simple addition, since the appropriate parsimony routines are in place?