One of the immediate-early oncogenes was named "jun" by Curran, who discovered it. It was named after his technician June Sonnenberg. (I knew her - before she got famous!)
There were many tyrosine kinases that were discovered through homology cloning. Many many many. One of them was named "jak", which was an acronym for "just another kinase". It turned out to be a rather important one.
So the only convention is that the gene is lower case and consists of three letters "fos" (or, hmmm, c-fos actually) and the protein, or enzyme, is capitalized: Fos.
But that was during the wise guy heyday of gene naming. Genes were named short and sweet so you could fit lots of names into drawings. Lots of genes have long names. And there is a long long big and very serious treatise on gene naming convenctions: http://www.gem.re.kr/CMR2/db_assignmentextver2.shtml
http://www.ciliate.org/help/gene_naming.shtml
The names are one thing when it is in the laboratory, very informal, and then it gets more and more formal as you register in GenBank and publish so you can all be sure you are talking about the same thing. But still quite anarchistic.