RE: [h-e-w] EmacsW32, gnuserv, pathes in .emacs

[Drew Adams]

    >The key requirement would be to somehow point out clearly which are the
    >things you really should learn first - distinguishing those from other
    >interesting things to learn, which are not so essential. Make
    it clear to
    >users - label the routes (like ski slopes?).

    Hmm, but as you say later, that quickly becomes a matter of taste,
    doesn't it?

No, I didn't say that what a user should learn first is a matter of taste.

When I spoke of taste and preference, I spoke of UIs. I said that UI
behavior and appearance, in particular, is often subject to user taste. Even
in the realm of UI, there are such things as good UI design and behavior.
The point was not that one shouldn't respect principles of good UI design,
but that users will always have their own idea of what's right for a UI
(whether they're right or wrong!).

This is not special to UI design, but it is stronger for UI design than some
other aspects of software: everyone thinks s?he is an expert, and that his
or her personal opinion represents good UI design. A similar thing happens
for documentation, BTW - everyone thinks s?he knows what's best for the
documentation. This happens, I think, because it is easy to have an opinion
on these things without being an expert in the subject. We tend to think
that a subject that we feel some familiarity with is simple, and that we
know as much as the experts. If we know nothing about some area, on the
other hand, we tend to trust the views of experts. We're of course wrong on
both counts: 1) experts know their area better than non-experts, and 2) they
are sometimes not to be trusted (simply because they are experts, in any
case).

Wrt what a user should learn first (the point you mentioned), a similar
principle does apply, however (even if I didn't say that). Users will have
their own preferences wrt learning. *And* we should try to point out what,
in general, are the best things to learn first. These things are relative,
and it's a question of guiding, not forcing users.

You do the best you can to provide a good UI, and you do the best you can to
point out what are the important things to learn first. And then you also
let users make their own choices, and you show them how to choose.

    I note that the OP mentioned "rectangle editing" in his
    third bullet. Would that have been one of your top few?

Mine personally? Or what the Emacs developers decide collectively, with user
input, and after discussion/debate?

    > I have never, ever, ever seen anyone use Emacs as is, out of
    > the box. Do you use only emacs -q? I'd be willing to bet on
    > that one.

       But let's face it, many of those key bindings -- the core ones --
    date back to dumb terminals on which you might not even have arrow
    keys, and 'ctrl' was the only key modifier. You had to (e.g.) press
    esc, shift-greater to jump to the end of a buffer.

Yes, I remember well. You're right that there is a legacy, but Emacs
developers constantly revisit the design and debate things like this (there
are different points of view among the developers too).

The reason some default bindings are what they are is because of a desire to
make Emacs usable out of the box on a multitude of platforms. That's why I
said that in some ways vanilla Emacs aims for the lowest common denominator.
And it *should* do so, even, sometimes, at the risk of not pleasing a
majority, who might have much more than the lowest common denominator.

There are many examples of this. For Emacs 22, for instance, global
font-lock mode will be enabled by default, for the first time. Previously,
it was considered that too many users might not have color available. The
point is that Emacs is not a slave to its legacy, and its designers do
actively revisit and debate its default settings.

       Before emacs I used an early PC-DOS editor which, for all its many
    limitations, turned out to be uncommonly well thought out. Surprise!
    The more I began learning about emacs, the more I realized the
    authors of my previous favorite editor had obviously set out to write
    a "baby emacs" suitable for the then-severely-underpowered PC.
       BUT -- since they knew every user had function keys, cursor keys,
    page keys, home/end keys, etc. etc., their default key bindings were
    comfortable: common tasks took ONE finger. I quickly became familiar
    with their conventions.

They were likely targeting one family of keyboards (PC). Emacs targets the
world of keyboards and beyond.

       So the very first thing I tried to do with real emacs was to
    borrow a .emacs file in which I could customize the key bindings to
    mimic what my fingers already knew.... And
    as you say, maybe starting from scratch with a brand-new version of
    emacs would be considerably easier than what I went through.

That was Eli who said that, but I agree with him on that, in general.

       Okay, that's a very long-winded way of supporting your position
    that customization is important. It IS the reason I chose emacs, and
    it's why I happily stay with it.

Of course - me too. Emacs *is* customization. If you're not customizing,
then you're not emacsing.