Per Abrahamsen writes:
> It depends. I think the *default* customize settings (such as in
`defface'
> forms) should be overridden with X resources. On the other hand,
> hand-changed customize settings (such as with `M-x customize-face') should
> override X resources.
Per> I'm not sure that is conceptually right. X resources gives the ability to
Per> make "per display" customizations, while customize only support
Per> customizations based on some broad display categories. The most specific
Per> should win. E.g. you may want to select different font sizes for
Per> different sized monitors , something customize cannot know about.
When faced with an existing specifier-like mechanism[1] we often
forget that the problems they solve is not really restricted to their domain.
What's interesting in the X resources mechanism is not that you can specify
on a per-display basis, but that for such or such display, you actually know
that you'll get a BW or color, big or small screen etc.
However, even if X provides a mechanism to cope with this variety of
displays, those problems are not specific to X. For instance, some terminals
support colors, some don't. We could also support the background pixmap
resource in Eterm.
But now, there's Xt and we're tempted to use it. It's a shame because
it's limited, and it makes it harder for us to think of the design of a more
generic specification model. What's missing now is a generic way of saying "I
want this size, this look and this color. If I happen to work on BW, let's do
this..."
In other words, IMHO, if with X you can manage to get the size you
want on all screens, but not with Custom, let's not conclude that X should
override Custom. Let's conclude that the user needs a higher way of setting
the font size, and X resources will be a "backend" for this model when running
under X.
I'm not saying this can be done in a few days :-)
Footnotes:
[1] The X resources mechanism is such, in the sense that you can "specify" on
a per-display, per-langage, per-screen ... basis.
--
/ / _ _ Didier Verna
http://www.inf.enst.fr/~verna/
- / / - / / /_/ / E.N.S.T. INF C201.1 mailto:vernaļ¼ inf.enst.fr
/_/ / /_/ / /__ / 46 rue Barrault Tel. (33) 01 45 81 73 46
75634 Paris cedex 13 Fax. (33) 01 45 81 31 19