A place for discussion of the potential adoption of a binomial nomenclature system for virus species names.
My preference would be to use #3 with a freeform name following the genus name. Using longer, more complex, species name such as Panicum streak mastrevirus was fine when only a “few” viral genomes were characterized (typically from cell-cultured isolates) and their cellular host and symptoms/disease signs were well known (and most often the reason they were studied). Now what we know of the biology of most viral genomes characterized using metagenomics is often extremely limited or non-existent (particularly for genomes from environmental rather than biological samples). The actual host and whether the virus is pathogenic is information often acquired later or for many viruses not at all. Indeed some information would be lost by changing name of species like Panicum streak mastrevirus to Mastrevirus panicum or Mastrevirus streak or Mastrevirus paniciumstreak but the phylogenetic information that it belong to the Mastrevirus genus would remain and an epithet can carry a lot of information. Those who study this or other viruses whose name would change would likley rapidly find out what the new binomial name of their favorite virus is.
The objection that information held within current virus species name would be lost using a binomial name is correct only for the shrinking proportion of viral species whose current name include more biological information that can be contained in a binomial name with a freeform epithet. A freeform epithet would have the advantage that species like Carnivore protoparvovirus 2 could become Protoparvovirus carnivore2. Species names like Cardiovirus A or Enterovirus B could stay the same. Another advantage of binomial classification is its simplicity for naming the torrent of “new” viral species currently being characterized. Assuming the genome can be classified into an existing genus the first name would be obvious and only an epithet needed to identify the species which could be any word even a made up word like tylcND1 which provide a lot of room for expected future expansion.
One issue would be for genomes that do not fall into an existing genus. In these cases a new genus name would have to be proposed followed by an epithet. Of course changing some well-established names could be disturbing, For example Human immunodeficiency virus 1 could become Lentivirus HIV1. Likely the resistance will be strong.