Re: delete-selection-mode (was: Put scroll-bar on right by defaultonUNIX.)

[Alan Mackenzie <acm@muc.de>]

Hi, Drew

On Thu, Mar 18, 2010 at 02:54:29PM -0700, Drew Adams wrote:
> > I am by no means unique - there will
> > certainly be lots of other people who suffer this feature likewise;
> > there're another one or two on this mailing list.  The degree of
> > suffering d-s-m inflicts on us far outweighs the slight increase in
> > convenience for you.

> Just say no. Just turn it off.

We're talking about the DEFAULT version here.  Would you now please
address the point.  That the change you want will impose a lot of pain
on lots of people, albeit that they may be in a minority.

> The discussion is about the _default_ behavior. Are you suggesting (a)
> that *most* new users already suffer from this problem outside Emacs
> and (b) they therefore need the Emacs default to be different from the
> outside behavior?

I'm suggesting that (a) a LOT of people suffer from this; and (b) they
would welcome Emacs being different.  (Emacs IS different in many ways.)

> > > 99.999% (no, no proof; just a guess) of computer users out there use
> > > this "risky" behavior everyday, all day long, without exploding (and
> > > without Emac's powerful undo as a remedy).

I've just spoken to my sister, an "ordinary" computer user.  She says
she normally uses the <delete> key after marking text before typing
further.  She also gets annoyed "every now and then" when marked text
gets accidentally deleted by typing, though "it's not too bad" if
there's an undo key sequence.

Unwanted deletion of text happens for me, it happens for David, it
happens for Miles.  Why do you not see this as a serious problem?

> > I use Emacs because it is (or rather, was) a stateless editor,

> Yes, OK, it is less modal than vi. More precisely, it has modes all
> over the place, instead of just 2 (or is it 3) modes. But Emacs is far
> from stateless.

STATES, not Modes.  Emacs WAS a stateless editor.  If you're in Foo
Mode, any key sequence did the same action always (Modulo deliberate
commands like C-s).  With t-m-m that no longer holds.  Should d-s-m be
made default it will be even less so.  I hold that this diminution of
statelessness is a Bad Thing.

> > d-s-m adds in yet one more frivolous state-dependent behaviour.

> Why frivolous?

Because inessential.  You and a few others, so as to save a single key
easy press (<del> or C-w) want to heap massive inconvenience on others.
In your personal way of working, you don't suffer from unwanted
deletion.  Others do.  What's so difficult about explicitly deleting
text in the region as opposed to it happening as a side effect?

> > Even with transient-mark-mode, you can still (currently) depend
> > on `self-insert-command' to just work. With d-s-m you can't.

> Sure you can. `self-insert-command' works fine with d-s-mode. You can
> depend on things acting the way they are documented, in a consistent
> way. That way is *different* from when d-s-mode is not enabled, but it
> is no less dependable.
 
`self-insert-command' will have state dependent behaviour.  It won't
JUST work anymore - it will have side effects.


> d-s-mode gives you a replace feature when the region is active, but it
> doesn't prevent you from having an inactive region and using it in
> other ways.

d-s-m makes an active region a fragile region.  It is this fragility
which causes all the pain.  Please address this issue.


> > How about addressing the question as put?  Is there any evidence
> > whatsoever for the intrinsic goodness of d-s-m?

> Using d-s-mode and not using it are about the same in terms of
> advantage/disadvantage, other things being equal: you need to hit an
> extra key in each to be able to get the behavior that the other gives
> you directly. With d-s-mode, the extra key is C-g (or C-u to prevent);
> without d-s-mode, the extra key is C-w (or `delete-region'/DEL). From
> this point of view, it's a toss-up.

That is mere hand waving, not evidence.  By evidence, I meant some sort
of study or research.  I take it you know of none.  I certainly don't.

> 2. Outside Emacs (Yes, Virginia, there is a world outside Emacs),
> type-to-replace is the rule for selections. Using the same rule as the
> default in Emacs helps both old users (only one behavior) and,
> especially, new users.

Who says that people outside Emacs use this rule much?  My sister
doesn't.


> > Hitting C-w is simple, hitting <del> is obvious even to newbies, and
> > doesn't make any noise.

> If the bell is the problem, we could perhaps silence it for this use.

What, more state?  No thanks!  Either silence it or not.  I'd say
silence it altogether.


> > > > One reason people might have come to Emacs is to escape 
> > > > the (to them) deity-awful key sequences they've been forced
> > > > to use up to now.

> > > That's an amazing statement, Alan. I've never heard anyone 
> > > claim that people come to Emacs because the key sequences
> > > they use elsewhere are too difficult.

> > Not "too difficult" but "deity-awful".  You do understand that
> > distinction, I hope?

> No. What did you mean exactly?

One which is "too difficult" is one which is difficult to use.
Ctrl-Alt-Delete would be difficult for a one-handed person.  Holding
down <PageUp> to go to BOB would be ghod-awful, as would C-x M-q C-w.
For kill-word, I'd call C-S-<right> <delete> ghod-awful, certainly when
compared with M-d.

> What is the salient characteristic of the keys outside Emacs that you
> think people complain about? "God-awful" might mean something concrete
> to you here, but it doesn't to me. Just which keys do you think they
> complain about? What God-awful keys outside Emacs make them come
> running inside?

C-f (for find) which drops a dialogue box over half of your text, for
example.  C-S-<right> <delete> for kill word.

> Did you happen to learn Emacs after using computers all day long for
> years, selecting and typing text to replace the selection? That's the
> case for folks nowadays. This is not 1985 or even 1995.

I've been using computers all day long for ~30 years; Emacs for ~12
years.  I've been assaulted by "typing replaces selection" for perhaps
the last 10 years or so.

> IIRC, you don't use a mouse (much, if at all), correct? And I'd guess
> you didn't use a mouse before you came to Emacs either. That is so
> different from 99.999999% of the world nowadays that it makes you miss
> the point, I fear, about _their_ learning Emacs.

Not at all - Emacs can wean them off their dependence on the mouse.

> It's not about you, Alan. And it's not about me. I turned on d-s-mode
> decades ago. I don't want the default change for myself. I want it for
> newbies, in particular.

It is about you, Drew.  I don't think you're looking outside your own
work habits enough to see that one size doesn't fit all.  d-s-m causes
distress; to David, to Miles, to me, to my sister, and undoubtedly to
countless others out there.

> I also think that some other oldbies will find it useful if they give
> it a chance. I'm struck by the number of oldbies, including RMS,
> who've made it clear in this very thread that they are not really
> familiar with d-s-mode. To any who are open, I say, "Try it; you might
> like it."

There's not too many familiar with BDSM either (and I'm not talking
about the licence here ;-).  Is that a good reason to try it?

-- 
Alan Mackenzie (Nuremberg, Germany).