Well, we have names for that: "horse" and "donkey". I get that having formal names is useful, and basing those names on ancestral characteristics seems sensibl

Well, we have names for that: "horse" and "donkey". I get that having formal names is useful, and basing those names on ancestral characteristics seems sensible. But it doesn't seem to take into account that this is a moving target: everything is evolving, and populations are not going to stay the same over time. Trying to classify every living thing into neat categories is like herding a million very slow cats. Especially with palaeontology, where the timespans are large enough for evolution to really matter.