For nucleotide, we use genomic.fa as the root (goofy, I know) and peptide.fa for protein. For some genomes, we also have ESTs and genes. These databases would clash if we followed the same naming scheme.
So dare I suggest two changes?
MH: can you update names so that genomic is "genomic.*" (ie genomic.nhr, etc) and peptide.* for blastp?
NdlC: can you update the search CGI and template to reflect this change?