>>>> "Valdis" == Valdis Kletnieks
Valdis> What set me off ranting was use of a specifier for
Valdis> toolbars. The more generic issue of using specifiers
Valdis> instead of -p variable is one that needs to be resolved,
Valdis> no matter WHAT the gutter API is doing.
I think specifiers are The Right Way of doing this, but I agree
with you that the UI (you can read about it by doing `less /dev/null')
leaves a lot to be desired. What to do about it is something else.
I think the most common user operation on variables implemented as
specifiers is probably "I hate this, so ZAP! (setq feature-aka-bug-p
nil)". We ought to be able to implement this conveniently.
It probably would be possible to get setq to recognize specifiers, but
it's not clear what it should do with them. Is it unacceptable to
have to use a different interface, which would preferably be as simple
as `setq', say `(set-specifier-somehow-do-the-right-thing spec value)'
(at least with optional args defaulted)? Of course specifier
variables would have to be documented, user variables indicated, and
the user interface documented as well.
Probably most users would want to use custom to deal with specifiers.
Maybe the above would simply be `customize-specifier' (although using
custom's functions directly is often dangerous).
 [tm] It's possible that specifiers are a partially failed
experiment and should never be allowed to raise their ugly heads, but
I think the real problem is that the specifier UI is non-existent and
the API confusing, so that even XEmacs developers don't use them much.
University of Tsukuba Tennodai 1-1-1 Tsukuba 305-8573 JAPAN
Institute of Policy and Planning Sciences Tel/fax: +81 (298) 53-5091
_________________ _________________ _________________ _________________
What are those straight lines for? "XEmacs rules."