Sub Ubi

I hate it when you spend half a day trying to get something to compile, only to have it “magically” work when you show it to someone else.

To finish off the technical report I’m writing, I really wanted to be able to download a running version of a tree-edit algorithm, and use that in my prototyping. I think I need to implement one of the algorithms, and be done with it.

I would rather like to do this without having to bother a paper author with questions about their algorithm. I must admit that I’ve learned something new: algorithmics papers are not typically eloquent. I have routinely encountered bits of papers where I have to interpret what I think the author meant if I’m going to do an implementation of my own.

If everyone provided the core algorithm in some concise, functional-esque language (say, Scheme), the world would be a much better place. (Read: pseudocode is always better than a plain-English description of the algorithm.)