Originally Posted by tore

I agree. I keep feeling there's a huge snag somewhere, but all I can see at the moment are problems that look/sound like they can be overcome, but that's no different from and perhaps even slightly better than for the other species definitions.

Comparing whole genomes would need some smart algorithms and so on so you'd need some clever programmers and mathematicians. However, even if there's a problem with an algorithm used to define species, at least you can just improve the algorithm and run it again. Although one should definetly try and find one best way to do it, you could easily define species in a number of different ways with something like this. For example, one could use different algorithms or perhaps restrict analysis to only certain bits of DNA (f.ex mitochondrial DNA) and so on.