This is the mail archive of the
guile@sourceware.cygnus.com
mailing list for the Guile project.
Re: hook? primitive
- To: "Greg Badros" <gjb@cs.washington.edu>
- Subject: Re: hook? primitive
- From: "Marisha Ray & Neil Jerram" <mpriz@dircon.co.uk>
- Date: Thu, 9 Sep 1999 08:55:36 +0100
- Cc: <guile@sourceware.cygnus.com>, "Carl R. Witty" <cwitty@newtonlabs.com>
From: Greg Badros <gjb@cs.washington.edu>
>That doesn't mean that you should have *less* information about a hooks
>intended procedure type. Ideally the hook would be able to report more
>information about the intended kind of procedure.
Isn't that rather like saying that all integer-type variables should somehow
know what sort of integer they are, such as a guile version number or the
number of hairs on a dog's head? In languages like Scheme and C, this sort
of semantic information is usually described implicitly by
- documentation
- context
- variable and procedure names
- knowledge in the programmers' heads.
Why make a special case for hooks?
Surely, at the point in SCWM where you are using, say, pointer-motion-hook,
you know the type of procedures that can be added to this hook?
Neil