The question of "what is a species" made me go a bit insane. I really wanted a unique identifier for each "species", but kept running into edge cases like species that changed scientific names (so you'd need name + date to resolve it). Occasionally people would refer to species by common name, but a common name can resolve to dozens of different scientific names. I don't think people realize how difficult and awkward it is to build software that deals with ambiguous entities.
This is a tangent, but the US Fish and Wildlife service has a cool tool for exploring endangered species [0]. I had a few issues with the data, but was overall impressed with how easy it was to access and how deep it goes (there are gigabytes of shapefiles representing endangered species boundaries offered in zip files). Preble's meadow jumping mouse [1] had a range so complex that its shapefile was 250 MB, the complete set of shapefiles covering all species' ranges was 2.5 GB--this single species of mouse took up 10% of this! For comparison, the Topeka Shiner's range shapefile was only 147 KB.
[0] - https://ecos.fws.gov/ecp/report/adhoc-creator?catalogId=spec...