Biology would be great, but in many instances the viruses currently cannot be cultured, and of course many instances where hosts have not been cultured. Yet, we can definitively state that a particular uncultured viruses is infecting a specific hosts. In many cases, ERVs and EVEs can be identified.

The taxonomic landscape has become extremely complicated and too dynamic. That said, there are many instances in which viruses that are only known from sequences are seen repeatedly, are clearly infecting their hosts, and should be classified. I'm not sure of a tractable resolution.