Re: face-remapping patch

[David Kastrup <dak@gnu.org>]

Richard M Stallman <rms@gnu.org> writes:

>     As I said: specifiers have the advantage of being able to
>     define/differentiate behavior even in the case where the locus is not
>     established yet.
>
> In what practical case is this an advantage?

Being able to define different parameters for different display types,
for example.  We have this for face definitions already.

Try
M-x customize-face RET highlight RET

then click on "State/For all kinds of displays".  You get what amounts
to a specifier of different faces that are "not yet" established.

Now for multi-tty support, on different tty's different faces will look
different.

This can be used in XEmacs for a unified approach to offering
display-dependent face, image, toolbar, menubar selections.

Which is a good idea when you are taking multi-tty support seriously
(which also might mean connecting from different computers to your local
Emacs session and having it open a terminal/frame on your external
terminal).  In practice, it is rather hard to beat application
programmers to actually provide all this versatility and work with it.

In theory, XEmacs could get along without buffer-local variables, using
specifiers instead.  In practice, enough programmers are already scared
off the areas in XEmacs which require the use of specifiers.

x-local variables seem less scary, but you can't use them (like in
customizations) before the locale x even exists.

In practice, uninstantiated specifiers might not be needed to get
manipulated outside of the customize interface in most cases.  So I
don't see we should create something like that with a life of its own.

-- 
David Kastrup, Kriemhildstr. 15, 44793 Bochum