Viruses are obligate intracellular parasites found in all living organisms and they were initially differentiated by the type of host they infected. As increasing knowledge of the structure of virus particles and of the biochemical and biological properties of viruses became available, viruses were classified in taxa known as orders, on the basis of the type and strandedness of their nucleic acids and presence of a reverse transcription process, and in family taxa on the basis of their particle structure, genome organization and virus replication strategy. Lower taxa such as genera and species were more difficult to establish because many additional criteria are needed for demarcating them, for instance natural host range, cell and tissue tropism, pathogenicity, vector transmission, small genome differences, antigenic properties etc. In spite of this, in 2016, a total of 3704 different virus species had already been established. With the recent advent of high-throughput sequencing and metagenomic approaches, tens of thousands of previously unrecorded viral sequences unrelated to any known viral genome were found to be present in a variety of marine and biological environments. Only a tiny fraction of these new sequences correspond to genomes of known virus species and no information is available regarding the host and biological properties of the unknown viruses harbouring these genome sequences. Recently, a group of 26 viral taxonomists convened a workshop to examine the possibility of incorporating these viral metagenomic sequences in the existing official virus classification system. Although viruses are currently classified using a combination of their chemical, structural, biological and genetic properties, virologists increasingly tend to rely mostly or exclusively on nucleotide genome sequences for assigning viruses to different taxa. The present review will discuss some of the problems that will be encountered when attempts are made to create viral species and genus taxa solely on the basis of genome nucleotide sequences.