The smallest of the small

The topic of a "minimal bacteriophage" came up in conservation with my lab's PI and postdoc yesterday. I got philosophical about it: what is a phage other than a genome, some structural proteins, and the means to enter bacterial cells? If we began removing items from a phage genome piece-by-piece, at what point would the genome cease to code for a phage? The questions really seem to concern how much genetic information is really necessary to self-propagate. 

There are some incredibly small phages out there, of course. Phage MS2 is a single-stranded RNA phage which infects E. coli and similar species. Its genome is just under 3600 nucleotides and contains just four genes. It was the first full genome to be sequenced - that small size made it a great candidate. It also seems to be used as a model for viruses in water quality studies.

So is MS2 a minimal bacteriophage? At the very least, its single capsid protein self-assembles into capsids and can even be used to encapsulate a protein of choice. If a phage needs to have a capsid then MS2 certainly passes the test. The other MS2 genes also provide basic functions: attachment, replication, and lysis. I'm guessing that the only way to get smaller would be to optimize or completely lose one of those functions (except for attachment, as it still needs to stick to something on its host surface), but would it still be a virus?