> AFAIK, only Kawa and MIT provide pathnames as a built-in, though there
> may be external packages that provide them on other systems.
> MIT is essentially CL in semantics, whereas Kawa is mostly about URIs.

I wonder whether they're ready for standardization, then.  On the
other hand, I certainly do like having the ability to manipulate
pathnames as something more than strings.

> That they are basically the same thing, and that URLs are essentially
> a mostly-upward-compatible extension.  Distinct parsers are needed
> because an URL is technically a valid Posix pathname, though a very
> unlikely one.

Perhaps, but one rarely uses them in the same context.  It seems much
less confusing to keep them separate.