Closed jeromekelleher closed 10 months ago
Personally, I think the idea of locally unary nodes is the main thing that makes people think that the local trees don't encode the entire ARG. When someone says "local trees" they think of each tree as only consisting of nodes that are branch points. If we lose the locally unary nodes, I can see that there isn't enough information to reconstruct the ARG (i.e. including recombination nodes).
Merging an update to the local trees and the "unary" nodes section. It's a bit more formal now, but since there was lots of confusion, I think this can't be avoided.
Hopefully some basic questions about why we bother with formal definitions etc have been answered.
A work in progress on the local trees section, trying to spell out why we think using a combinatorial encoding is useful and important.
Needs more work, but actually this seems quite an important insight...