PHYLIP can also generate all possible trees if you make a data set
that has no information (either zero sites or sequences that are all
"?" states), and then use (say) DNAPENNY to do a branch-and-bound search for
all best trees. But since I put a 100-tree limit on output you do need to
recompile to increase that constant first.
By the way in the paper by Mary Kuhner and I in Molecular Biology and
Evolution 11: 459-468 (1994) we describe two metric distances between
trees, the Robinson and Foulds dT measure and our own concoction, the
Branch Score, which is the square of a metric distance. The former
depends only on the topology, the latter also on branch lengths.
Joe Felsenstein, Dept. of Genetics, Univ. of Washington, Seattle, WA 98195
Internet: joe at genetics.washington.edu (IP No. 126.96.36.199)