I still think it would be very desirable to tie the short and long options together. Eg.
options = [('verbose', 'v'), ('quiet', 'q'), ('thingy', None), (None, 'x') ('output=', 'o:')] opts, args = getopt_or_die (options, usage, ...)
Im not convinced this is worth it. I only use "long options" when I have too many, or a few obscure ones. I have never have "-v" synonymous for "--verbose" - why bother? I know I would never type the later:-)
The existing mechanism still handles this quite well - the standard "if opt==blah:" simply becomes "if opt in [...]:" - no real drag.
Plus its less change for reasonable reward - handy enough I may actually add command-line handling as I create each little test/util script :-)
What say anyone else? Go with my "little change", Gregs "only slightly more change" or "don't worry about it"?