Saying you would recommend against use without saying why you would and then
saying that alone is a good enough reason is not much of an argument, actually.
So, why would you recommend against its usage? Kebab-case is quite popular for
program names in the Unix world. Fully 20% of my `/usr/bin` programs have a
`'-'` in them (about 3X the number that use underscore). (These are not
programs named by me.) Also popular is having the name of the program match the
name of the source file up to an extension. So, if you mean to say this 75%
super majority of dash vs underscore program namers is so wrong-headed, what
exactly have they got wrong?
As for "dealing with code" that uses it, use in library/module names as opposed
to program names is unlikely to be common. That said, tossing backticks on
there doesn't sound very much of a burden..It might be less onerous in some
more system-wide sense than having a "library version" and a "program version"
or symlinks or other kinds of aliasing some people might concoct to work around
the problem.
I think having a `{.out:.}` pragma to let a program set its own (default,
overridable on the `nim c` command-line) name is a good step, too. That might
solve most people's problems. It would still seem lamely inflexible that module
names are more restricted than other identifier names (procs, vars, etc.).
