Your proposal is a bit better than the original, although it still duplicates the puts (imagine a variant where you’d want to handle I/O errors), and some will be bothered by the fact that the same unchanging condition is being retested in each loop iteration (the compiler may even warn about it).

But still, I don’t see why you wouldn’t first name what you want to output before starting the outputting. If anything, I’d place the whole output loop in a separate function and have two calls to that function. Nevertheless, it’s even better to express in code the fact that the program doesn’t want to make a distinction between a literal “y” and an argument “y”, by consolidating them into the same variable.

Another way to do this would be to have a static default argument array containing the “y”, and for example having:

    if (argc <= 1) { argv = default_argv; }
    for (;;) { puts(argv[1]); }

This would make explicit the fact thst the argument-less invocation is merely a shortcut for an invocation with an argument and doesn’t otherwise provide any new or different behavior.

Though I think the separate variable (what) is clearly preferable.