I had the dubious pleasure of being assigned to revive some NASA FORTRAN code from 1978 during grad school. FORTRAN 77 was too new, so it was written in FORTRAN 66 style. With vendor extensions, naturally. Fortunately this wasn't CFD code. It was an order of magnitude less complicated. Anyhow, nobody had compiled this thing in 15 years. It took about a month for me to figure out how to make it work, because it wasn't obvious that the garbage I was getting out of it was because I was compiling it wrong. Eventually I found the right flags for gfortran, and it just worked.
So yes, if you count grad students, people are working on FORTRAN for their bread.
