Re: PATCH: Handle symbols in messages

[Jim Blandy <jimb@redhat.com>]

Doug Evans <dje@transmeta.com> writes:
> fwiw, while string-concat might not clearly specify the operation
> (though I don't have a problem with it), `concat' seems too generic
> (or at least not string-specific enough in the way that the rest of
> the string functions are so named).
> 
> string-map is to map,
> as string-list is to list,
> as string-<mumble> is to concat.
> 
> concat mightn't make sense outside of strings (and thus why have the
> string- prefix ...), but why preclude symbol-concat? (etc.)
> 
> How about string-concat?

Thanks for picking at this.  I had thought I was naming it after a
Common Lisp function, but now that I look it up I see I was getting my
dialects confused.  (Emacs Lisp has a concat function that does type
conversions on its arguments.)

How about string/symbol-append?  Given that there's no good precedent
for 'concat' after all, and that concat has another meaning in CGEN
.cpu files, and that the distinction between appending and
concatenating is unclear, perhaps we should stick to the verb Scheme
generally uses for this kind of operation, and simply name the types
we're handling.