C-d deleting region considered harmful

C-d deleting region considered harmful

[Miles Bader]

I notice that various deletion commands will now delete an active
region.  The most obviously useful to newbies is DEL (aka backspace),
as that's what other platforms use, and I don't have any problems with
that (and I might even use it occasionally)

However, the effect of "C-d" deleting the active region has been
driving me nuts for quite a while.  I accidentally end up deleting
something I didn't intend to about 10-15 times a day... usually I
notice and am annoyed but can immediately hit undo, but occasionally I
don't immediately notice, and end up being very very confused until I
reconstruct what happened.

I thought I might get used to it after a while, but so far, despite
the frequency with which it happens, it's every bit as annoying as it
was when it first started happening.

Is there any particular _reason_ for C-d to have this effect?  C-d is
not a binding used by other platforms, so it has no obvious utility as
newb-bait, and it's no more convenient than C-w.  It seems almost
entirely pointless.

I propose removing this functionality from C-d.  If mac/windows users
want to delete a region, they can use DEL (backspace).

Thanks,

-Miles

-- 
Bigot, n. One who is obstinately and zealously attached to an opinion that
you do not entertain.

Re: C-d deleting region considered harmful

[Miles Bader]

BTW, I realize that DEL/backspace might theoretically have the same
problem, but it's also a clearly more useful binding, and for whatever
reason doesn't seem to be problematic in practice for me.

-miles

-- 
Custard, n. A vile concoction produced by a malevolent conspiracy of the hen,
the cow, and the cook.

Re: C-d deleting region considered harmful

[Stefan Monnier]

> Is there any particular _reason_ for C-d to have this effect?  C-d is

I don't know what was Chong's motivation for it, but one reason i that
the `delete' key is often remapped to C-d.


        Stefan

Re: C-d deleting region considered harmful

[Eli Zaretskii]

> From: Miles Bader <miles@gnu.org>
> Date: Sat, 18 Sep 2010 10:49:19 +0900
> 
> I propose removing this functionality from C-d.  If mac/windows users
> want to delete a region, they can use DEL (backspace).

There's no DEL key on Windows, only <Delete>, which is remapped to
C-d.  Maybe the command should distinguish whether it was run by C-d
or <Delete>, and behave differently wrt the active region.

Deleting an active region with Backspace is a bad idea, IMO: that's
not how other apps behave, while this feature exists to be compatible
with other apps.

Re: C-d deleting region considered harmful

[David Kastrup]

Eli Zaretskii <eliz@gnu.org> writes:

>> From: Miles Bader <miles@gnu.org>
>> Date: Sat, 18 Sep 2010 10:49:19 +0900
>> 
>> I propose removing this functionality from C-d.  If mac/windows users
>> want to delete a region, they can use DEL (backspace).
>
> There's no DEL key on Windows, only <Delete>, which is remapped to
> C-d.  Maybe the command should distinguish whether it was run by C-d
> or <Delete>, and behave differently wrt the active region.
>
> Deleting an active region with Backspace is a bad idea, IMO: that's
> not how other apps behave, while this feature exists to be compatible
> with other apps.

Just tried it in the location bar of firefox.  Backspace certainly does
delete an active region there.

-- 
David Kastrup

Re: C-d deleting region considered harmful

[Eli Zaretskii]

> From: David Kastrup <dak@gnu.org>
> Date: Sat, 18 Sep 2010 11:45:25 +0200
> 
> > Deleting an active region with Backspace is a bad idea, IMO: that's
> > not how other apps behave, while this feature exists to be compatible
> > with other apps.
> 
> Just tried it in the location bar of firefox.  Backspace certainly does
> delete an active region there.

Firefox does a lot of things that others don't.

Re: C-d deleting region considered harmful

[Eli Zaretskii]

> Date: Sat, 18 Sep 2010 12:06:08 +0200
> From: Eli Zaretskii <eliz@gnu.org>
> Cc: emacs-devel@gnu.org
> 
> > From: David Kastrup <dak@gnu.org>
> > Date: Sat, 18 Sep 2010 11:45:25 +0200
> > 
> > > Deleting an active region with Backspace is a bad idea, IMO: that's
> > > not how other apps behave, while this feature exists to be compatible
> > > with other apps.
> > 
> > Just tried it in the location bar of firefox.  Backspace certainly does
> > delete an active region there.
> 
> Firefox does a lot of things that others don't.

And in any case, hitting <Delete> is a customary op, so removing it is
not a good idea.

P.S.  Note that I don't use active regions at all, so my personal
habits will not be affected either way.

Re: C-d deleting region considered harmful

[dhruva]

> From: Eli Zaretskii <eliz@gnu.org>
> To: dak@gnu.org, emacs-devel@gnu.org
> Date: Sat, 18 Sep 2010 12:15:23 +0200
> Subject: Re: C-d deleting region considered harmful
>> Date: Sat, 18 Sep 2010 12:06:08 +0200
>> From: Eli Zaretskii <eliz@gnu.org>
>> Cc: emacs-devel@gnu.org
>>
>> > From: David Kastrup <dak@gnu.org>
>> > Date: Sat, 18 Sep 2010 11:45:25 +0200
>> >
>> > > Deleting an active region with Backspace is a bad idea, IMO: that's
>> > > not how other apps behave, while this feature exists to be compatible
>> > > with other apps.
>> >
>> > Just tried it in the location bar of firefox.  Backspace certainly does
>> > delete an active region there.
>>
>> Firefox does a lot of things that others don't.
>
> And in any case, hitting <Delete> is a customary op, so removing it is
> not a good idea.
>
> P.S.  Note that I don't use active regions at all, so my personal
> habits will not be affected either way.

For deleting a a portion of text, I always use the shift->Arrow
selection to mark it and hit backspace to
delete the region. Since shift->arrow key does not work when I use
"emacs -nw" mode, I struggle. To
summarize, I heavily depend on ability to delete regions using the
backspace. I just hope it does not change. It may be a relic of being
a windows user but old habits die hard!

-dhruva

Re: C-d deleting region considered harmful

[Sebastian Rose]

Eli Zaretskii <eliz@gnu.org> writes:
>> Date: Sat, 18 Sep 2010 12:06:08 +0200
>> From: Eli Zaretskii <eliz@gnu.org>
>> Cc: emacs-devel@gnu.org
>> 
>> > From: David Kastrup <dak@gnu.org>
>> > Date: Sat, 18 Sep 2010 11:45:25 +0200
>> > 
>> > > Deleting an active region with Backspace is a bad idea, IMO: that's
>> > > not how other apps behave, while this feature exists to be compatible
>> > > with other apps.
>> > 
>> > Just tried it in the location bar of firefox.  Backspace certainly does
>> > delete an active region there.
>> 
>> Firefox does a lot of things that others don't.
>
> And in any case, hitting <Delete> is a customary op, so removing it is
> not a good idea.
>
> P.S.  Note that I don't use active regions at all, so my personal
> habits will not be affected either way.


This change has removed a reaonable distinction between deleting a
character (C-d) and deleting a region (C-w) which was a good thing.

Yet another bad change.

Re: C-d deleting region considered harmful

[Miles Bader]

Eli Zaretskii <eliz@gnu.org> writes:
>> I propose removing this functionality from C-d.  If mac/windows users
>> want to delete a region, they can use DEL (backspace).
>
> There's no DEL key on Windows, only <Delete>, which is remapped to
> C-d.  Maybe the command should distinguish whether it was run by C-d
> or <Delete>, and behave differently wrt the active region.

"DEL" is emacs-speak for "backspace" on typical PC keyboards.

> Deleting an active region with Backspace is a bad idea, IMO: that's
> not how other apps behave, while this feature exists to be compatible
> with other apps.

Hmm, are you sure?  I thought Backspace was _the_ traditional
delete-a-region key on macs and windows -- it certainly is on the mac
anyway (since forever), and I'm pretty sure that's also what I use to
delete stuff when I use a windows system... (since windows derived most
of its basic keybindings from the mac, it would be ery surprising if it
weren't the same)

-Miles

-- 
gravity a demanding master ... soft soft snow

Re: C-d deleting region considered harmful

[Miles Bader]

dhruva <dhruvakm@gmail.com> writes:
> To summarize, I heavily depend on ability to delete regions using the
> backspace. I just hope it does not change. It may be a relic of being
> a windows user but old habits die hard!

Note that I'm complaining about "C-d", not backspace.

-Miles

-- 
Arrest, v. Formally to detain one accused of unusualness.

RE: C-d deleting region considered harmful

[Drew Adams]

> However, the effect of "C-d" deleting the active region has been
> driving me nuts for quite a while.  I accidentally end up deleting
> something I didn't intend to about 10-15 times a day... usually I
> notice and am annoyed but can immediately hit undo, but occasionally I
> don't immediately notice, and end up being very very confused until I
> reconstruct what happened.
> 
> I thought I might get used to it after a while, but so far, despite
> the frequency with which it happens, it's every bit as annoying as it
> was when it first started happening.
> 
> Is there any particular _reason_ for C-d to have this effect?  C-d is
> not a binding used by other platforms, so it has no obvious utility as
> newb-bait, and it's no more convenient than C-w.  It seems almost
> entirely pointless.
> 
> I propose removing this functionality from C-d.  If mac/windows users
> want to delete a region, they can use DEL (backspace).

100% agreement.  This is truly a bad idea.  C-d (and <delete>) in Emacs should
do what it always has done: delete the next char.

[I did not realize that C-d had been changed this way, since I still don't use
Emacs 24 (while waiting for the selection stuff and other bugs to die down).  I
see now that this was introduced in thread "Updated proposal for DEL to delete
active region".  I did not see that it was part of that proposal at the time.
Most of the discussion there was about the commands and their names and mappings
and other problems about breaking compatibility.  This additional change slipped
right by me.  I wouldn't have missed it if C-d had been mentioned in the Subject
line instead of DEL.  DEL already deletes the active region in delete-selection
mode (which I use), but for C-d to do that is a real change.]

Please reverse this change.  This is a terrible design decision.

RE: C-d deleting region considered harmful

[Drew Adams]

> BTW, I realize that DEL/backspace might theoretically have the same
> problem, but it's also a clearly more useful binding, and for whatever
> reason doesn't seem to be problematic in practice for me.

Yes, and DEL/backspace has always deleted the active region in delete-selection
mode.

RE: C-d deleting region considered harmful

[Drew Adams]

> > Is there any particular _reason_ for C-d to have this 
> > effect?
> 
> I don't know what was Chong's motivation for it, but one reason i that
> the `delete' key is often remapped to C-d.

That's not a reason.  That's not a reason to change the behavior of either/both
of them.

People maybe often call you "Stef" (dunno).  That's not a reason why both "Stef"
and "Stefan" should suddenly start referring to a completely different
person/behavior whenever the moon is full. 

RE: C-d deleting region considered harmful

[Drew Adams]

> > I propose removing this functionality from C-d.  If 
> > mac/windows users want to delete a region, they can use DEL (backspace).
> 
> There's no DEL key on Windows, only <Delete>, which is remapped to
> C-d.  Maybe the command should distinguish whether it was run by C-d
> or <Delete>, and behave differently wrt the active region.

No.  Both <delete> and C-d, whether mapped together or not, should do what they
have always done in Emacs: delete the next char. Whether the region is active or
not.

> Deleting an active region with Backspace is a bad idea, IMO: that's
> not how other apps behave, while this feature exists to be compatible
> with other apps.

Huh? That's how the Backspace key behaves in other apps on Windows, AFAICT.
(Well, "active region" is only an Emacs thing, but for Windows the selection is
deleted when you hit Backspace.)

Re: C-d deleting region considered harmful

[Bastien]

"Drew Adams" <drew.adams@oracle.com> writes:

>> I propose removing this functionality from C-d.  If mac/windows users
>> want to delete a region, they can use DEL (backspace).
>
> 100% agreement.  This is truly a bad idea.  C-d (and <delete>) in Emacs should
> do what it always has done: delete the next char.

+1

-- 
 Bastien

RE: C-d deleting region considered harmful

[Drew Adams]

> > > Deleting an active region with Backspace is a bad idea, 
> > > IMO: that's not how other apps behave, while this feature
> > > exists to be compatible with other apps.
> > 
> > Just tried it in the location bar of firefox.  Backspace 
> > certainly does delete an active region there.
> 
> Firefox does a lot of things that others don't.

And IE and Outlook and Word and Windows Explorer and...

Hitting Backspace in most apps on Windows deletes the selection.

RE: C-d deleting region considered harmful

[Drew Adams]

> For deleting a a portion of text, I always use the shift->Arrow
> selection to mark it and hit backspace to
> delete the region. Since shift->arrow key does not work when I use
> "emacs -nw" mode, I struggle. To
> summarize, I heavily depend on ability to delete regions using the
> backspace. I just hope it does not change. It may be a relic of being
> a windows user but old habits die hard!

Me too.  But this thread is not about stopping Backspace from deleting the
active region.  It is about stopping C-d and the <delete> key from doing that.

RE: C-d deleting region considered harmful

[Drew Adams]

> This change has removed a reaonable distinction between deleting a
> character (C-d) and deleting a region (C-w) which was a good thing.
> 
> Yet another bad change.

+1, including for "yet another".

Emacs had the best selection/region behavior around, especially if a user chose
to use delete-selection mode.  For months now selection/region behavior has been
in the process of stumbling "improvement", leading to lots of bug reports and a
fair amount of grumbling.

Personally, I'm just waiting for this work-in-progress to die down and the final
customization settings to be made known that let users restore the previous,
sane selection/region behavior.

There are so many ways Emacs could _really_ be improved.  Why is it so hard to
learn "If it ain't broke don't fix it"?

Re: C-d deleting region considered harmful

[Christoph]

On 9/17/2010 7:49 PM, Miles Bader wrote:

> However, the effect of "C-d" deleting the active region has been
> driving me nuts for quite a while.

I am trying to understand the real problem here. I personally think the 
new behavior is more intuitive, but anyway. If you mark an active 
region, why exactly do you press C-d? Wouldn't you expect the command 
after selecting the active region to be executed on the active region?

Christoph

Re: C-d deleting region considered harmful

[Stefan Monnier]

> No.  Both <delete> and C-d, whether mapped together or not, should do
> what they have always done in Emacs: delete the next char. Whether the
> region is active or not.

Thanks.  But we're discussing here what to do with C-d: for `delete' we
already decided that it should delete an active region (by default).
Whether you agree with this `delete' behavior is irrelevant.


        Stefan

Re: C-d deleting region considered harmful

[Christoph]

On 9/18/2010 6:02 AM, Sebastian Rose wrote:

> This change has removed a reaonable distinction between deleting a
> character (C-d) and deleting a region (C-w) which was a good thing.
>
> Yet another bad change.

Actually, C-d runs delete-char (or in 24.0.50 delete-forward-char) and 
C-w runs kill-region. I thought the difference between delete and kill 
was that the latter puts the killed section on the kill-ring, no? I 
would definitely appreciate a distinctive behavior like that.

Btw, in 23.2 (on Windows), how do I delete (not kill) an active region?

C-Space, C-f a couple of times to select word, then neither Backspace, 
nor Delete actually delete the active region. I can only kill it with C-w.

Christoph

Re: C-d deleting region considered harmful

[Eli Zaretskii]

> Date: Sat, 18 Sep 2010 10:14:30 -0600
> From: Christoph <cschol2112@googlemail.com>
> CC: sebastian_rose@gmx.de, Eli Zaretskii <eliz@gnu.org>, dak@gnu.org
> 
> Btw, in 23.2 (on Windows), how do I delete (not kill) an active region?

"M-x delete-region RET".

Re: C-d deleting region considered harmful

[Christoph]

On 9/18/2010 10:20 AM, Eli Zaretskii wrote:

>> Btw, in 23.2 (on Windows), how do I delete (not kill) an active region?
>
> "M-x delete-region RET".

Hmm, that's quite the effort required for a simple thing like that. I 
like C-d a lot better. ;)

Christoph

Re: C-d deleting region considered harmful

[Christoph]

On 9/18/2010 9:05 AM, Drew Adams wrote:

> No.  Both<delete>  and C-d, whether mapped together or not, should do what they
> have always done in Emacs: delete the next char. Whether the region is active or
> not.

As for C-d and its command delete-forward-char, the documentation states:

To disable this, set `delete-active-region' to nil.

Imho, the default behavior is more consistent than the old behavior. C-d 
without an active region deletes the character, with an active region 
deletes the region. Makes sense to me.

Christoph

Re: C-d deleting region considered harmful

[David Kastrup]

dhruva <dhruvakm@gmail.com> writes:

> For deleting a a portion of text, I always use the shift->Arrow
> selection to mark it and hit backspace to
> delete the region. Since shift->arrow key does not work when I use
> "emacs -nw" mode, I struggle. To
> summarize, I heavily depend on ability to delete regions using the
> backspace. I just hope it does not change. It may be a relic of being
> a windows user but old habits die hard!

If the region has been selected moving backwards, it would appear
natural to have DEL/C-d delete the region.  It is probably too clever to
let DEL/C-d/Backspace not delete the active region only if it "faces the
other way".  Though it would likely cure Miles' problem.

-- 
David Kastrup

Re: C-d deleting region considered harmful

[Miles Bader]

Christoph <cschol2112@googlemail.com> writes:

>> However, the effect of "C-d" deleting the active region has been
>> driving me nuts for quite a while.
>
> I am trying to understand the real problem here. I personally think the
> new behavior is more intuitive, but anyway. If you mark an active
> region, why exactly do you press C-d? Wouldn't you expect the command
> after selecting the active region to be executed on the active region?

I press C-d to delete the character after the cursor -- the command is
called "delete-forward-char" after all.  Adding a new meaning when the
region is activated may or may not be an extra convenience, but it's
not unambiguously natural behavior.

It's not uncommon to have the region be activated somewhat
inadvertently; this is generally harmless, but since the traditional
behavior in emacs is to simply deactivate the region when one hits
C-d, I've apparently become used to hitting C-d to perform its
traditional functionality, even when the region is active.  Because
C-d is a very low-level command, this sort of thing usually occurs in
a quick sequence of commands, not as a carefully considered event
where I think deeply about the meaning and consequence of each
keystroke before executing it.

As I mentioned, DEL/backspace theoretically suffer the same issue, but
don't seem to be a problem for me in practice; maybe that's because
I've also become used to the "extra functionality" in that case due to
using mac/windows software.

Moreover, adding this new functionality to DEL/backspace has
undeniable utility, because many many mac/windows users have that
particular usage hardwired into their fingers.  This is _not_ true of
C-d.

-miles

-- 
People who are more than casually interested in computers should have at
least some idea of what the underlying hardware is like.  Otherwise the
programs they write will be pretty weird.  -- Donald Knuth

RE: C-d deleting region considered harmful

[Drew Adams]

> > However, the effect of "C-d" deleting the active region has been
> > driving me nuts for quite a while.
> 
> I am trying to understand the real problem here...
> If you mark an active region, why exactly do you press C-d?

To delete a char, the char after point.

> Wouldn't you expect the command after selecting the active
> region to be executed on the active region?

No, I would expect to act as it always has, on the char after point.

RE: C-d deleting region considered harmful

[Drew Adams]

> As for C-d and its command delete-forward-char, the 
> documentation states:
>  To disable this, set `delete-active-region' to nil.

That also prevents DEL (Backspace) from deleting the active region.

But at least setting it to nil does not prevent DEL from working correctly in
delete-selection mode.  So yes, setting it to nil is a solution for me, since I
use delete-selection mode.

I still think this "feature" of having C-d and <delete> delete the active region
is misguided for Emacs.  But as long as it doesn't affect delete-selection mode
I cannot complain for my own use.

I do complain for Emacs, however, that the default value of the option changes
the default behavior.  If those behind this "improvement" want to let users
optionally delete the region with C-d and <delete>, then OK, provide that as an
option.  But do not make it the default.

Keep the default behavior as it was, and let users try the new Wunder behavior
optionally, if they so wish.

RE: C-d deleting region considered harmful

[Drew Adams]

> > No.  Both <delete> and C-d, whether mapped together or not, 
> > should do what they have always done in Emacs: delete the
> > next char.  Whether the region is active or not.
> 
> Thanks.  But we're discussing here what to do with C-d:

Yes, we are.  And I'm adding <delete> to the discussion, which is typically
mapped to C-d.

> for `delete' we already decided that it should delete an
> active region (by default).

I see.  Where/when was that decided?

That's too bad.  C-d and <delete> should just delete the next char.

> Whether you agree with this `delete' behavior is irrelevant.

Anything I say you might consider irrelevant, whether I agree or disagree with
you.  That's your right.  I still voice my agreement or disagreement.

Voicing disagreement to your decisions is often ineffectual in terms of getting
you to reconsider, it is true.  That is not a reason to keep silent when a
choice is bad.

Re: C-d deleting region considered harmful

[Lars Magne Ingebrigtsen]

Miles Bader <miles@gnu.org> writes:

> As I mentioned, DEL/backspace theoretically suffer the same issue, but
> don't seem to be a problem for me in practice; maybe that's because
> I've also become used to the "extra functionality" in that case due to
> using mac/windows software.

The "DEL deleting the region" behaviour was something that was driving
me insane.  I usually jump around a lot with `C-x C-x' in buffers, and
then I may want to delete something, and I use `DEL' to delete stuff
with.

So I just switched off `transient-mark-mode', which is something I
suspect most Emacs old-timers will be more comfortable with.

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen

Re: C-d deleting region considered harmful

[Lars Magne Ingebrigtsen]

"Drew Adams" <drew.adams@oracle.com> writes:

> Keep the default behavior as it was, and let users try the new Wunder
> behavior optionally, if they so wish.

While I have some sympathy for that train of thought (there's nothing as
enraging as a trusted tool suddenly doing something unexpected), that
means that all new users have to suffer forever with an old-timey
behaviour, even though the new behaviour might make a lot more sense for
new users.

After all, experienced Emacs users know (or should know) that they can
twiddle some variables to get any old behaviour back.

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen

Re: C-d deleting region considered harmful

[Miles Bader]

On Sun, Sep 19, 2010 at 3:44 AM, Drew Adams <drew.adams@oracle.com> wrote:
>> for `delete' we already decided that it should delete an
>> active region (by default).
>
> I see.  Where/when was that decided?
>
> That's too bad.  C-d and <delete> should just delete the next char.

I personally don't care what happens to <delete> (I never, ever, use
it), but to be fair, I think there is a case for it having the "delete
region" functionality:  (1) it seems to do that on "other" software,
and (2) It's the only key on most keyboards that's actually labelled
"delete", so it's not surprising that naive users might expect it to
delete the region they just carefully selected...

[and again, neither is true of "C-d"...]

-Miles


-- 
Cat is power.  Cat is peace.

Re: C-d deleting region considered harmful

[Miles Bader]

Lars Magne Ingebrigtsen <larsi@gnus.org> writes:
> The "DEL deleting the region" behaviour was something that was driving
> me insane.  I usually jump around a lot with `C-x C-x' in buffers, and
> then I may want to delete something, and I use `DEL' to delete stuff
> with.
>
> So I just switched off `transient-mark-mode', which is something I
> suspect most Emacs old-timers will be more comfortable with.

I really like transient-mark-mode actually, and generally find the
"active region functionality" of most commands both useful and
intuitive.

C-d is not one of those commands, of course; really I think it's simply
too low-level a command, and the extra functionality not really a
natural extension of the basic functionality.

Hmm, it might be cute if the "active region" stuff was actually handled
by the key-binding lookup mechanism!

(global-set-key [active-region ?\C-d]
                (lambda (n beg end)
                  (interactive "p\nr")
                  (if (yes-or-no-p
                       "Do you really, really, want to delete the region?!?!?")
                     (kill-region beg end)
                    (delete-forward-char n))))

-Miles

-- 
Is it true that nothing can be known?  If so how do we know this?  -Woody Allen

RE: C-d deleting region considered harmful

[Drew Adams]

> > But we're discussing here what to do with C-d:
> > for `delete' we already decided that it should delete an
> > active region (by default).
> 
> I see.  Where/when was that decided?

I cannot find such a decision in the mailing list.  AFAICT, the <delete> key
wasn't even mentioned in Yidong's proposal ("Updated proposal for DEL to delete
active region"), except to remove it from `mouse-region-delete-keys'.

His proposal addressed the commands, not the keys, putting the <delete> and
<deletechar> keys squarely in the same boat as `C-d' (and `DEL').

So why do you apparently consider that <delete> has been decided and cannot be
discussed, whereas the question is still open for `C-d'?  Where/when did you
decide that the <delete> key, but not (yet) `C-d', should delete the active
region?

RE: C-d deleting region considered harmful

[Drew Adams]

> > Keep the default behavior as it was, and let users try the 
> > new Wunder behavior optionally, if they so wish.
> 
> While I have some sympathy for that train of thought (there's 
> nothing as enraging as a trusted tool suddenly doing something
> unexpected), that means that all new users have to suffer forever
> with an old-timey behaviour, even though the new behaviour might
> make a lot more sense for new users.

Bzzzt!  No, but thanks for playing.

All new users will NOT have to suffer forever.  In fact, NO new users will have
to suffer forever.  New users, like old users, can discover options and change
their settings if they like.

The default behavior should be the behavior that we think is best out of the
box.  I don't happen to agree that the default behavior for C-d (and <delete>)
should be changed.

> After all, experienced Emacs users know (or should know) that
> they can twiddle some variables to get any old behaviour back.

Would that it were (always). ;-)

But I am hoping and expecting that the promises of new settings to get back all
of the pre-24 selection/region behavior will be respected.  Still waiting for
the dust to settle on that one.

Re: C-d deleting region considered harmful

[Lars Magne Ingebrigtsen]

"Drew Adams" <drew.adams@oracle.com> writes:

> All new users will NOT have to suffer forever.  In fact, NO new users
> will have to suffer forever.  New users, like old users, can discover
> options and change their settings if they like.

If new users sees a tool behaving strangely, they say "how odd", and
then go on to the next tool.

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen

Re: C-d deleting region considered harmful

[Chong Yidong]

Stefan Monnier <monnier@IRO.UMontreal.CA> writes:

> I don't know what was Chong's motivation for it, but one reason is
> that the `delete' key is often remapped to C-d.

Indeed.  I suppose we could get away with mapping C-d to delete-char and
[delete] to delete-forward-char.  The main disadvantage would be the
breaking of the equivalence of C-d and DEL.

RE: C-d deleting region considered harmful

[Drew Adams]

> If the region has been selected moving backwards, it would appear
> natural to have DEL/C-d delete the region.

If you want the direction to come into play here, then I would think you would
be proposing this for C-d (only) when the region is selected backwards and DEL
(only) when selected forwards.  Yes, one could argue that that is natural in
some sense.  But having DEL delete in both directions is inconsistent with
wanting C-d to be directional.

> It is probably too clever to let DEL/C-d/Backspace not delete
> the active region only if it "faces the other way".  Though it
> would likely cure Miles' problem.

FWIW, in Windows both DEL (Backspace) and <delete> (Delete) delete the active
selection, but C-d does lots of different things depending on the context (it is
more likely to delete some other object - e.g. an entire mail message or
application window).

RE: C-d deleting region considered harmful

[Drew Adams]

> > All new users will NOT have to suffer forever.  In fact, NO 
> > new users will have to suffer forever.  New users, like old
> > users, can discover options and change their settings if they like.
> 
> If new users sees a tool behaving strangely, they say "how odd", and
> then go on to the next tool.

Some yes, but not all.  You said "all new users".

Following that logic, ALL of the Emacs default UI should be exactly what all new
users expect - nothing more.  (Let's ignore here the fact that new users do not
_all_ expect the same things.)  We should then make no attempt to move newbies
in what we think is a better direction.

That has not been the approach for designing Emacs (although some periodically
do argue for such a knee-jerk recipe).  Instead, each feature and its options
are examined from various aspects and a judgment is made as to whether and how
much to risk surprising newbies but also guide them toward something better.

IOW, the devil is in the details: each potential default behavior is discussed
on its merits, including its effect on newbies.  There is no hard-and-fast rule
that everything must reflect what most (not to mention all) newbies are already
used to.  Not surprising newbies is only one consideration, albeit an important
one.

Re: C-d deleting region considered harmful

[Sebastian Rose]

Christoph <cschol2112@googlemail.com> writes:
> On 9/18/2010 10:20 AM, Eli Zaretskii wrote:
>
>>> Btw, in 23.2 (on Windows), how do I delete (not kill) an active region?
>>
>> "M-x delete-region RET".
>
> Hmm, that's quite the effort required for a simple thing like that. I like C-d a
> lot better. ;)

Until you delete a big bunch of text by accident.


You could (re-)bind keys to make something "simple" like deleting a
whole bunch of text for good the default for you.


The current situation is, we have to rebind more and more keys to do
something reasonable in many, many cases.  That's worse.


  Sebastian



PS: You'll find more information about the `kill-ring', `kill-ring-max',
`browse-kill-ring' and other usefull Emacs tools in the Emacs manual.

Re: C-d deleting region considered harmful

[Christoph]

On 9/18/2010 1:41 PM, Sebastian Rose wrote:

> You could (re-)bind keys to make something "simple" like deleting a
> whole bunch of text for good the default for you.

Why wouldn't something simple like that be available with a simple key 
combination out of the box?

> The current situation is, we have to rebind more and more keys to do
> something reasonable in many, many cases.  That's worse.

I don't understand what you mean.

Christoph

Re: C-d deleting region considered harmful

[Sebastian Rose]

Christoph <cschol2112@googlemail.com> writes:
> On 9/18/2010 9:05 AM, Drew Adams wrote:
>
>> No.  Both<delete>  and C-d, whether mapped together or not, should do what they
>> have always done in Emacs: delete the next char. Whether the region is active or
>> not.
>
> As for C-d and its command delete-forward-char, the documentation states:
>
> To disable this, set `delete-active-region' to nil.
>
> Imho, the default behavior is more consistent than the old behavior. C-d 
> without an active region deletes the character, with an active region deletes
> the region. Makes sense to me.

Consistent with what???

The old behaviour was _clever_ !



Imagine you have created an active region. Once you have done that, you
need to be really carefull how to proceed (slow down).



If you now decide to keep the region, you cannot go on with the next
natural editing command.  Instead, you must deactivate the region first,
or the text is gone (not for good, but see below).  (extra key press,
extra slot for remembering the region => slow down)

In those oh soo good "normal" applications you want Emacs to be
consistent with, any editing command will now replace the region.  Wich
is the course Emacs is going, obviously.  C-d does this already and the
rest will follow, since the newbies seem to tell Emacs what to do (to
find new users... bad intention --- this will make Emacs more "normal",
not better).

Note, that deliting just one character is a simple editing command, just
as adding one character.

Any of these simple editing commands can delete a whole bunch of text in
"normal" applications.  You might be used to that behaviour because it's
"normal", but I'm sure it caused you some headache from time to time :)


In Emacs, if you decide to delete the region, it's just pressing a `w'
instead of a `d' (or any other simple editing command).

In neither case, the region will be lost --- in Emacs.

Simple, isn't it?



Now imagine, for some reason, you have deleted the region.  A few
minutes later you recognize, it was a mistake.

Now, in "normal" apps, you press "UNDO" many many times, until you get
your region back.  You lose all the work done from deleting the region
onwards until this point in time.  This slows down editing again.  This
is stupid.  This is anoying even.


That's why we need to distinguish between commands that act on region
(e.g. `C-w') and ALL other simple editing commands, including `C-d'.




As most people here, I rarely use backspace or delete.  They're awkward
to type.  But defining them differently is a surprise, true enough.
This means the decision made was a mistake.





Sebastian

Re: C-d deleting region considered harmful

[Christoph]

On 9/18/2010 12:40 PM, Miles Bader wrote:

> I press C-d to delete the character after the cursor -- the command is
> called "delete-forward-char" after all.

In 23.2 you use C-d to delete the character UNDER the cursor. Unless you 
have a different definition of cursor than I do.

delete-forward-char does not make much sense to me either. From the 
documentation the only difference to delete-char is said behavior or 
killing the active region. Technically, it's a delete backwards.

I do like the behavior though.

> It's not uncommon to have the region be activated somewhat
> inadvertently;

I usually activate the region to perform some kind of a command on it. 
If I decide that I need to delete the character under the cursor instead 
of deleting to entire region, is it too strange to hit C-g first, to 
deactivate the region and then press C-d to delete the character? That 
seems like a pretty normal workflow to me. Unless you are expecting 
implicit behavior like deactivating the region before deleting the 
character. As an Emacs user of 2 years (not 200) I wouldn't necessarily 
expect implicit (or traditional) behavior like that. ;)

> Moreover, adding this new functionality to DEL/backspace has
> undeniable utility, because many many mac/windows users have that
> particular usage hardwired into their fingers.  This is _not_ true of
> C-d.

Absolutely true. C-d, to me, is a convenient alias for the <delete> key, 
which on most keyboards is inefficiently out of reach. Every single 
Windows application that I just pulled up from Explorer, to Firefox, 
Chrome, Open Office or whichever, behaves the same: pressing <delete> on 
selected text deletes the entire selection. That, of course, does not 
mean Emacs must do the same, but just serves as an example that there is 
many many users who would expect this to work like they are used to.

Christoph

Re: C-d deleting region considered harmful

[Sebastian Rose]

> The default behavior should be the behavior that we think is best out of the
> box.

THIS is reasonable.


>> After all, experienced Emacs users know (or should know) that
>> they can twiddle some variables to get any old behaviour back.

This is NOT reasonable at all and was said too many times.



BTW: New users are NO reason for a change.  It just pure virtual and
holds no truth at all -- it cannot.

Most "normal" applications are inferior to Emacs, that's why we all use
Emacs in the end.

So how can those "normal" applications rule Emacs ????


I for one do not want to hear that virtual "new users" argument any more
on a reasonable mailing list.  It is NO argument.






Technical advantages, reasonable behaviour and preferences ARE arguments
for changes or, in this case, against them.


Deleting regions through simple editing commands is NOT reasonable.
It slows down editing and is non-productive (some say dangerous).




Sebastian

Re: C-d deleting region considered harmful

[Eli Zaretskii]

> From: Lars Magne Ingebrigtsen <larsi@gnus.org>
> Date: Sat, 18 Sep 2010 20:47:18 +0200
> 
> So I just switched off `transient-mark-mode', which is something I
> suspect most Emacs old-timers will be more comfortable with.

Here's one such old-timer.

Re: C-d deleting region considered harmful

[Richard Stallman]

    I notice that various deletion commands will now delete an active
    region.  The most obviously useful to newbies is DEL (aka backspace),
    as that's what other platforms use, and I don't have any problems with
    that (and I might even use it occasionally)

Did people go ahead with that change?  Such a change should not be made
without polling the users first.

I tried it for a few days and found it rather painful.

Re: C-d deleting region considered harmful

[David Kastrup]

Christoph <cschol2112@googlemail.com> writes:

> On 9/18/2010 12:40 PM, Miles Bader wrote:
>
>> I press C-d to delete the character after the cursor -- the command is
>> called "delete-forward-char" after all.
>
> In 23.2 you use C-d to delete the character UNDER the cursor. Unless
> you have a different definition of cursor than I do.

Likely.  Emacs has the cursor _between_ characters.  vi has the cursor
_on_ characters.  So vi has commands for inserting before and after a
character, before and after a line and so on.  Emacs has only one kind
of insertion, _at_ point.

-- 
David Kastrup

Re: C-d deleting region considered harmful

[David Kastrup]

Lars Magne Ingebrigtsen <larsi@gnus.org> writes:

> "Drew Adams" <drew.adams@oracle.com> writes:
>
>> All new users will NOT have to suffer forever.  In fact, NO new users
>> will have to suffer forever.  New users, like old users, can discover
>> options and change their settings if they like.
>
> If new users sees a tool behaving strangely, they say "how odd", and
> then go on to the next tool.

What's wrong with that?  We are not drug traffickers having anything to
gain by getting people hooked.

-- 
David Kastrup

Re: C-d deleting region considered harmful

[Lars Magne Ingebrigtsen]

David Kastrup <dak@gnu.org> writes:

> What's wrong with that?  We are not drug traffickers having anything to
> gain by getting people hooked.

Sure we are.  We're hoping people get addicted to Emacs so they start
contributing code to Emacs.

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen

Re: C-d deleting region considered harmful

[David Kastrup]

Lars Magne Ingebrigtsen <larsi@gnus.org> writes:

> David Kastrup <dak@gnu.org> writes:
>
>> What's wrong with that?  We are not drug traffickers having anything
>> to gain by getting people hooked.
>
> Sure we are.  We're hoping people get addicted to Emacs so they start
> contributing code to Emacs.

I don't think you can reasonably expect contributions from people who
would be abhorred at the thought of looking into a manual.

-- 
David Kastrup

RE: C-d deleting region considered harmful

[Drew Adams]

> Likely.  Emacs has the cursor _between_ characters.  vi has the cursor
> _on_ characters.  So vi has commands for inserting before and after a
> character, before and after a line and so on.  Emacs has only one kind
> of insertion, _at_ point.

Yes.  Maybe this helps too:

The cursor is a graphic indication of the text insertion position, aka point.
The position is unambiguous, no matter how the cursor might indicate that
position.

Typically, when the cursor shape is not narrow enough to show between two chars
it is displayed more or less on top of the char that follows the insertion
point.

A bar cursor is narrow enough to show between two chars.  A box cursor is wide
enough that it is shown in Emacs on top of the char after point.

Re: C-d deleting region considered harmful

[Christoph]

On 9/19/2010 6:40 AM, Drew Adams wrote:

>> Likely.  Emacs has the cursor _between_ characters.  vi has the cursor
>> _on_ characters.  So vi has commands for inserting before and after a
>> character, before and after a line and so on.  Emacs has only one kind
>> of insertion, _at_ point.
>
> Yes.  Maybe this helps too:
>
> The cursor is a graphic indication of the text insertion position, aka point.
> The position is unambiguous, no matter how the cursor might indicate that
> position.
>
> Typically, when the cursor shape is not narrow enough to show between two chars
> it is displayed more or less on top of the char that follows the insertion
> point.
>
> A bar cursor is narrow enough to show between two chars.  A box cursor is wide
> enough that it is shown in Emacs on top of the char after point.

I should have known better...thanks for the explanation.

Christoph

Re: C-d deleting region considered harmful

[Sebastian Rose]

Christoph <cschol2112@googlemail.com> writes:
> I usually activate the region to perform some kind of a command on it. If I
> decide that I need to delete the character under the cursor instead of deleting
> to entire region, is it too strange to hit C-g first, to deactivate the region
> and then press C-d to delete the character? That seems like a pretty normal
> workflow to me.


C-g is a superfluous extra keystroke in this case.

It will be (is) typed billions of times just because of a wrong
decision.


Sebastian

Re: C-d deleting region considered harmful

[Christoph]

On 9/18/2010 3:18 PM, Eli Zaretskii wrote:

>> So I just switched off `transient-mark-mode', which is something I
>> suspect most Emacs old-timers will be more comfortable with.
>
> Here's one such old-timer.

Out of curiosity, from old-timer to new-timer, what advantages does the 
traditional Emacs behavior over transient-mark-mode?

Christoph

Re: C-d deleting region considered harmful

[Eli Zaretskii]

> From: "Drew Adams" <drew.adams@oracle.com>
> Date: Sun, 19 Sep 2010 05:40:12 -0700
> Cc: 
> 
> The cursor is a graphic indication of the text insertion position, aka point.
> The position is unambiguous, no matter how the cursor might indicate that
> position.

Of course, this nice theory completely breaks down with bidirectional
text, at the directionality change boundaries.

Re: C-d deleting region considered harmful

[Lars Magne Ingebrigtsen]

Christoph <cschol2112@googlemail.com> writes:

> Out of curiosity, from old-timer to new-timer, what advantages does
> the traditional Emacs behavior over transient-mark-mode?

The advantage for me is that I don't keep deleting text that I don't
want to delete.  :-)

In addition, I don't have to stomp on `C-g' for every third command I do
to make the region go away and let me edit stuff as normal.

(I think `C-g' should only be used when you really want to get out of a
sticky situation -- not as part of the normal editing routine.)

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen

Re: C-d deleting region considered harmful

[Eli Zaretskii]

> Date: Sun, 19 Sep 2010 08:00:39 -0600
> From: Christoph <cschol2112@googlemail.com>
> CC: Eli Zaretskii <eliz@gnu.org>, larsi@gnus.org
> 
> On 9/18/2010 3:18 PM, Eli Zaretskii wrote:
> 
> >> So I just switched off `transient-mark-mode', which is something I
> >> suspect most Emacs old-timers will be more comfortable with.
> >
> > Here's one such old-timer.
> 
> Out of curiosity, from old-timer to new-timer, what advantages does the 
> traditional Emacs behavior over transient-mark-mode?

I use "C-x C-x" _a_lot_ as a kind of one-time bookmark (augmented by
"C-u C-SPC" when necessary), to jump quickly between two spots in a
buffer.  With this paradigm, the region is not really "region", it's
just a portion of text between 2 loci of interest.  Therefore, it
drives my nuts that Emacs paints the text in color each time I type
"C-x C-x", because I have no intention of marking the region, I just
want to jump.

OTOH, I always have C-w or M-w if I want to kill the region.

So I have a very good reason _not_ to use transient-mark-mode, and
none to use it.

Re: C-d deleting region considered harmful

[Sebastian Rose]

Christoph <cschol2112@googlemail.com> writes:
> On 9/18/2010 1:41 PM, Sebastian Rose wrote:
>
>> You could (re-)bind keys to make something "simple" like deleting a
>> whole bunch of text for good the default for you.
>
> Why wouldn't something simple like that be available with a simple key
> combination out of the box?


It is.  `C-w'


>> The current situation is, we have to rebind more and more keys to do
>> something reasonable in many, many cases.  That's worse.
>
> I don't understand what you mean.

It was stated in one of the mails before, that "oldtimers" could re-bind
keys to keep them doing, in this case, the sensible thing.  The subject
of this thread is just one out of several discussions and changes that
came up this year.




Sebastian

Re: C-d deleting region considered harmful

[Christoph]

On 9/19/2010 7:55 AM, Sebastian Rose wrote:

> C-g is a superfluous extra keystroke in this case.
> It will be (is) typed billions of times just because of a wrong
> decision.

Sebastian,
I get your point. You like the traditional way of Emacs' behavior and 
the more I read about it, the better I understand why.

However, there are no absolute truths when it comes to discussing 
efficiency in editing, since there are always personal preferences and 
habits involved. I might argue, that selecting text (with transient mark 
mode) and then replacing the text instantly with new text by inserting 
characters, like any other editor does, is a lot more efficient than 
having to kill the selection first (or let alone, delete with M-x 
delete-region!) and then inserting characters. You object, since that is 
not the way you are used to working.

Christoph

Re: C-d deleting region considered harmful

[Christoph]

On 9/19/2010 8:17 AM, Sebastian Rose wrote:

>>> You could (re-)bind keys to make something "simple" like deleting a
>>> whole bunch of text for good the default for you.
>>
>> Why wouldn't something simple like that be available with a simple key
>> combination out of the box?
>
> It is.  `C-w'

Changes my kill-ring, no?

> It was stated in one of the mails before, that "oldtimers" could re-bind
> keys to keep them doing, in this case, the sensible thing.  The subject
> of this thread is just one out of several discussions and changes that
> came up this year.

OK, understood. Thanks.

Christoph

Re: C-d deleting region considered harmful

[Sebastian Rose]

Christoph <cschol2112@googlemail.com> writes:
>>>> You could (re-)bind keys to make something "simple" like deleting a
>>>> whole bunch of text for good the default for you.
>>>
>>> Why wouldn't something simple like that be available with a simple key
>>> combination out of the box?
>>
>> It is.  `C-w'
>
> Changes my kill-ring, no?


Yes.

My kill-ring-max is set to 500

 ;)




Sebastian

Re: C-d deleting region considered harmful

[Chad Brown]

Question for the people who dislike transient-mark-mode and it's
related behaviors:

Do you find yourself making an unwanted `highlighted region' using any
other method than C-xC-x?  The potential candidates that come to mind
are `C-space and movement', shift-selection, and mouse-sweep.

I used to dislike transient-mark-mode (back when it was called
zmacs-region and I was using Epoch), and disabled it along with most
of the `chrome', but at one point I intentionally tried working with
all of the bells and whistles and found it helpful except when I was
using C-xC-x often when editing code.  I spend more time editing text
these days, and much less hoping around with C-xC-x, so I just live
with the occasional distraction, but perhaps a way to invert the ARG
of exchange-point-and-mark would let more emacs hackers enjoy the
benefits of t-m-m without being annoyed by the spurious flashing/C-g.

*Chad

Re: C-d deleting region considered harmful

[Wojciech Meyer]

Chad Brown <yandros@MIT.EDU> writes:

> Question for the people who dislike transient-mark-mode and it's
> related behaviors:
>
> Do you find yourself making an unwanted `highlighted region' using any
> other method than C-xC-x?  The potential candidates that come to mind
> are `C-space and movement', shift-selection, and mouse-sweep.
>
> I used to dislike transient-mark-mode (back when it was called
> zmacs-region and I was using Epoch), and disabled it along with most
> of the `chrome', but at one point I intentionally tried working with
> all of the bells and whistles and found it helpful except when I was
> using C-xC-x often when editing code.  I spend more time editing text
> these days, and much less hoping around with C-xC-x, so I just live
> with the occasional distraction, but perhaps a way to invert the ARG
> of exchange-point-and-mark would let more emacs hackers enjoy the
> benefits of t-m-m without being annoyed by the spurious flashing/C-g.

I find `transient-mark-mode' quite useful, some of the commands work in
a different way, maybe more intuitively (for instance `comment-dwim' or
some of the replace functions).

However I've been annoyed quite few times when it highlights the region
when I don't want it. C-x C-x being a primal example (yes, I understand
that it's a change in semantics).

So I think overall it is a good improvement for handling regions in
Emacs.

However, I would not like to see C-d killing region, as in my work flow
I quite often used it with selected region (most importantly it is
better to separate C-d and backspace). Also people who are aware of C-d
most likely are people who used Emacs for some time, in a contrast to
the users who just hit backspace. So the behaviour of having a choice,
`kill or not to kill' the region with a pressure on a intuitive default
behaviour for both groups of users seems to reasonable.

>
> *Chad

Cheers;
Wojciech

Re: C-d deleting region considered harmful

[Lars Magne Ingebrigtsen]

Wojciech Meyer <wojciech.meyer@googlemail.com> writes:

> However I've been annoyed quite few times when it highlights the region
> when I don't want it. C-x C-x being a primal example (yes, I understand
> that it's a change in semantics).

Perhaps `C-x C-x' is the prime annoyance factor here?  It certainly was
for me.  Perhaps it just shouldn't activate the transient mark mode?

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen

Re: C-d deleting region considered harmful

[Eli Zaretskii]

> From: Chad Brown <yandros@MIT.EDU>
> Date: Sun, 19 Sep 2010 11:00:50 -0700
> 
> Do you find yourself making an unwanted `highlighted region' using any
> other method than C-xC-x?

Only with "C-x C-x".  But that's one of the most frequent commands I
use.

Re: C-d deleting region considered harmful

[Alan Mackenzie]

'afternoon, Chad,

On Sun, Sep 19, 2010 at 11:00:50AM -0700, Chad Brown wrote:
> Question for the people who dislike transient-mark-mode and it's
> related behaviors:

> Do you find yourself making an unwanted `highlighted region' using any
> other method than C-xC-x?  The potential candidates that come to mind
> are `C-space and movement', shift-selection, and mouse-sweep.

No, never.  I disabled transient-mark-mode a split microsecond after it
was thrust upon me.  I think most hackers who "dislike" this will have
done likewise.  By the way, "dislike" doesn't come close to how I regard
transient-mark-mode; visceral revulsion is closer to the mark and point.

I still occasionally get this ghastly effect (e.g. when I start emacs -Q
for some testing reason).  Sometime I'll get around to hacking my copy of
the place where this default is set.

> I used to dislike transient-mark-mode (back when it was called
> zmacs-region and I was using Epoch), and disabled it along with most
> of the `chrome', but at one point I intentionally tried working with
> all of the bells and whistles and found it helpful except when I was
> using C-xC-x often when editing code.  I spend more time editing text
> these days, and much less hoping around with C-xC-x, so I just live
> with the occasional distraction, but perhaps a way to invert the ARG
> of exchange-point-and-mark would let more emacs hackers enjoy the
> benefits of t-m-m without being annoyed by the spurious flashing/C-g.

I'm sure t-m-m could be made nicer.  For a start, it's a bumbling
conflation of three things which ought to be independently setable,
namely (i) Highligting of the region; (ii) a form of narrowing to the
region; (iii) disablement of some region commands.

In fact, if I could disable (i), that intrusive dark blue obliteration of
my font-lock and hi-lock highlighting, I might even get to try t-m-m.
But then, many of the config variables are so badly named
("transient-mark-mode" ought to be "transient-region-highlighting" since
the mark is no longer transient by default - all region commands are by
default executable) it would put me off even just setting these newspeak
variables.

BTW, I vote for C-d keeping its "do one thing and do it well" binding of
`delete-char', even though I might have disqualified myself from voting.
When I have to use inferior software I detest that software marking
regions without my consent and thereby forcing my right hand away from
the home position to remove this frivolous marking on pain of losing the
entire region.

In fact, I think I predicted the current confusion when the decision to
make t-m-m a default was taken.

> *Chad

-- 
Alan Mackenzie (Nuremberg, Germany).

Re: C-d deleting region considered harmful

[David Kastrup]

Eli Zaretskii <eliz@gnu.org> writes:

>> From: "Drew Adams" <drew.adams@oracle.com>
>> Date: Sun, 19 Sep 2010 05:40:12 -0700
>> Cc: 
>> 
>> The cursor is a graphic indication of the text insertion position, aka point.
>> The position is unambiguous, no matter how the cursor might indicate that
>> position.
>
> Of course, this nice theory completely breaks down with bidirectional
> text, at the directionality change boundaries.

Well, when we are typing linearly, we don't want the block cursor to be
on the character we just typed, but ahead.

I just did C-h h C-x C-q, went into the Hebrew text and typed a few
characters (like x or ( ) since ( ) are not strictly L-R).

I got seasick.

-- 
David Kastrup

Re: C-d deleting region considered harmful

[Eli Zaretskii]

> From: David Kastrup <dak@gnu.org>
> Date: Sun, 19 Sep 2010 21:18:12 +0200
> 
> I just did C-h h C-x C-q, went into the Hebrew text and typed a few
> characters (like x or ( ) since ( ) are not strictly L-R).
> 
> I got seasick.

I hope your keyboard or input method mirror characters, or else you
will really become seasick.

Re: C-d deleting region considered harmful

[Christoph]

On 9/19/2010 8:46 AM, Sebastian Rose wrote:

>>> It is.  `C-w'
>>
>> Changes my kill-ring, no?
>
> Yes.
>
> My kill-ring-max is set to 500

I am not concerned about losing stuff but having to do more work to yank 
what I want. :)

Christoph

Re: C-d deleting region considered harmful

[Christoph]

On 9/19/2010 8:12 AM, Eli Zaretskii wrote:

> I use "C-x C-x" _a_lot_ as a kind of one-time bookmark (augmented by
> "C-u C-SPC" when necessary), to jump quickly between two spots in a
> buffer.  With this paradigm, the region is not really "region", it's
> just a portion of text between 2 loci of interest.  Therefore, it
> drives my nuts that Emacs paints the text in color each time I type
> "C-x C-x", because I have no intention of marking the region, I just
> want to jump.

I see the problem. Can't say I ever used C-x C-x like that. But that 
sounds like a pretty useful thing. :)

Thanks,
Christoph

Re: C-d deleting region considered harmful

[Sebastian Rose]

Chad Brown <yandros@MIT.EDU> writes:
> Question for the people who dislike transient-mark-mode and it's
> related behaviors:
>
> Do you find yourself making an unwanted `highlighted region' using any
> other method than C-xC-x?  The potential candidates that come to mind
> are `C-space and movement', shift-selection, and mouse-sweep.

No.

I never use `C-x C-x'.
I never use shift-selection.
I don't use the mouse to edit text either.

Re: C-d deleting region considered harmful

[Eli Zaretskii]

> Date: Sun, 19 Sep 2010 15:02:58 -0600
> From: Christoph <cschol2112@googlemail.com>
> Cc: emacs-devel@gnu.org
> 
> > My kill-ring-max is set to 500
> 
> I am not concerned about losing stuff but having to do more work to yank 
> what I want. :)

"Edit->Paste from Kill Menu" to the rescue.

Re: C-d deleting region considered harmful

[Sebastian Rose]

Christoph <cschol2112@googlemail.com> writes:
>>>> It is.  `C-w'
>>>
>>> Changes my kill-ring, no?
>>
>> Yes.
>>
>> My kill-ring-max is set to 500
>
> I am not concerned about losing stuff but having to do more work to yank what I
> want. :)

I suspect you're aware of `M-y', right?
The last kill is always on top of the kill-ring.
Keeping more than one thing for yanking is one of the unique strengths
of emacs. 


Sebastian

Re: C-d deleting region considered harmful

[Christoph]

On 9/19/2010 3:28 PM, Eli Zaretskii wrote:
>> Date: Sun, 19 Sep 2010 15:02:58 -0600
>> From: Christoph<cschol2112@googlemail.com>
>> Cc: emacs-devel@gnu.org
>>
>>> My kill-ring-max is set to 500
>>
>> I am not concerned about losing stuff but having to do more work to yank
>> what I want. :)
>
> "Edit->Paste from Kill Menu" to the rescue.

You want me to use the MOUSE???? ;)

Re: C-d deleting region considered harmful

[David De La Harpe Golden]

On 19/09/10 22:39, Sebastian Rose wrote:

> Keeping more than one thing for yanking is one of the unique strengths
> of emacs.
>

Once upon a time maybe. Clipboard histories are now fairly commonplace.

Re: C-d deleting region considered harmful

[Christoph]

On 9/19/2010 3:39 PM, Sebastian Rose wrote:

> I suspect you're aware of `M-y', right?
> The last kill is always on top of the kill-ring.
> Keeping more than one thing for yanking is one of the unique strengths
> of emacs.

Yes I am aware of that. I actually use browse-kill-ring for maximum 
convenience with the kill-ring. But I think you misunderstood: I want 
only things on the kill-ring that I put there. I don't want things to 
end up on the kill ring which weren't supposed to go there, i.e. deleted 
items. So, the solution of using C-w for deleting (not killing) a region 
is suboptimal, since it pollutes my kill-ring. I press M-y which (with 
browse-kill-ring) pops up a nice selection window and I have to sift 
through all of these deleted items instead of my nice and clean 
collection of kills.

Christoph

Re: C-d deleting region considered harmful

[Richard Stallman]

    Thanks.  But we're discussing here what to do with C-d: for `delete' we
    already decided that it should delete an active region (by default).
    Whether you agree with this `delete' behavior is irrelevant.

It won't bother anyone for `delete' to delete the region.  Making
`backspace' do so is the change that will bother lots of users.

Re: C-d deleting region considered harmful

[Lennart Borgman]

On Mon, Sep 20, 2010 at 2:16 AM, Richard Stallman <rms@gnu.org> wrote:
>    Thanks.  But we're discussing here what to do with C-d: for `delete' we
>    already decided that it should delete an active region (by default).
>    Whether you agree with this `delete' behavior is irrelevant.
>
> It won't bother anyone for `delete' to delete the region.  Making
> `backspace' do so is the change that will bother lots of users.

However a lot of other people also expect C-d to delete the visible
region. I think we really need to get compatibility of this kind
implemented. (I have some ideas for it, but I have a lack of time at
the moment so I have preferred not to say much about it.)

Re: C-d deleting region considered harmful

[Miles Bader]

Lars Magne Ingebrigtsen <larsi@gnus.org> writes:
>> However I've been annoyed quite few times when it highlights the region
>> when I don't want it. C-x C-x being a primal example (yes, I understand
>> that it's a change in semantics).
>
> Perhaps `C-x C-x' is the prime annoyance factor here?  It certainly was
> for me.  Perhaps it just shouldn't activate the transient mark mode?

Not a good idea, I think, since one of the primary uses of C-x C-x is to
re-activate the region!  [for me about 90% of the time I think...]

It'd be good to have an alternate, unbound-by-default variant that
didn't activate the mark though, to make rebinding easy for those that
prefer the no-activate behavior.

-Miles

-- 
/\ /\
(^.^)
(")")
*This is the cute kitty virus, please copy this into your sig so it can spread.

Re: C-d deleting region considered harmful

[Miles Bader]

Richard Stallman <rms@gnu.org> writes:
>     Thanks.  But we're discussing here what to do with C-d: for `delete' we
>     already decided that it should delete an active region (by default).
>     Whether you agree with this `delete' behavior is irrelevant.
>
> It won't bother anyone for `delete' to delete the region.  Making
> `backspace' do so is the change that will bother lots of users.

But it's also the only such change that really matters for
newbie-friendless.

[and again, for the record, I find "backspace deletes region" to be
absolutely no problem for me in practice.  It's C-d that sucks...]

-Miles

-- 
Friendship, n. A ship big enough to carry two in fair weather, but only one
in foul.

RE: C-d deleting region considered harmful

[Drew Adams]

> >    Thanks.  But we're discussing here what to do with C-d: 
> >    for `delete' we already decided that it should delete an
> >    active region (by default).  Whether you agree with this
> >    `delete' behavior is irrelevant.
> >
> > It won't bother anyone for `delete' to delete the region.  Making
> > `backspace' do so is the change that will bother lots of users.
> 
> However a lot of other people also expect C-d to delete the visible
> region.

Huh?  Where do you get that?  Who would expect that based on other applications
or platforms?  That is certainly not what C-d does on Windows, for instance.

> I think we really need to get compatibility of this kind
> implemented.

Compatibility?  With what?

> (I have some ideas for it, but I have a lack of time at
> the moment so I have preferred not to say much about it.)

Start by saying something to back up your preposterous claims that a lot of
people are used to C-d deleting the selection and that there is a
"compatibility" imperative for C-d to do so.

Re: C-d deleting region considered harmful

[David Kastrup]

Christoph <cschol2112@googlemail.com> writes:

> On 9/19/2010 3:28 PM, Eli Zaretskii wrote:
>>> Date: Sun, 19 Sep 2010 15:02:58 -0600
>>> From: Christoph<cschol2112@googlemail.com>
>>> Cc: emacs-devel@gnu.org
>>>
>>>> My kill-ring-max is set to 500
>>>
>>> I am not concerned about losing stuff but having to do more work to yank
>>> what I want. :)
>>
>> "Edit->Paste from Kill Menu" to the rescue.
>
> You want me to use the MOUSE???? ;)

You can get there using f10.  So the complaint should rather be that he
wants you to use your eye-hand-coordination.

-- 
David Kastrup

Re: C-d deleting region considered harmful

[David Kastrup]

Miles Bader <miles@gnu.org> writes:

> Lars Magne Ingebrigtsen <larsi@gnus.org> writes:
>>> However I've been annoyed quite few times when it highlights the region
>>> when I don't want it. C-x C-x being a primal example (yes, I understand
>>> that it's a change in semantics).
>>
>> Perhaps `C-x C-x' is the prime annoyance factor here?  It certainly was
>> for me.  Perhaps it just shouldn't activate the transient mark mode?
>
> Not a good idea, I think, since one of the primary uses of C-x C-x is to
> re-activate the region!  [for me about 90% of the time I think...]

If you just want to reactivate the region, the proper keysequence to do
only that and nothing else would be C-x C-x C-x C-x (otherwise you also
move point).  So one could make this reactivate only on second call
(like C-SPC does).  On the other hand, C-x C-x C-x C-x is often used for
checking "did I leave the mark just where intended?".

-- 
David Kastrup

Re: C-d deleting region considered harmful

[Lars Magne Ingebrigtsen]

Miles Bader <miles@gnu.org> writes:

> Not a good idea, I think, since one of the primary uses of C-x C-x is to
> re-activate the region!  [for me about 90% of the time I think...]

I don't even know what "re-activate the region" means, so it's not a
primary use for me.  :-)

People use `C-x C-x' for different purposes, which makes it more
complicated to satisfy everybody here, perhaps.

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen

Re: C-d deleting region considered harmful

[David Kastrup]

Lars Magne Ingebrigtsen <larsi@gnus.org> writes:

> Miles Bader <miles@gnu.org> writes:
>
>> Not a good idea, I think, since one of the primary uses of C-x C-x is to
>> re-activate the region!  [for me about 90% of the time I think...]
>
> I don't even know what "re-activate the region" means, so it's not a
> primary use for me.  :-)

Try C-u C-x C-x some time.

-- 
David Kastrup

Re: C-d deleting region considered harmful

[Bastien]

RMS suggested to poll the users for such issues.

This was also suggested many times when the list was discussing t-m-m to
be turned on by default.

There wasn't any poll.

A poll is never an easy thing to do: you need to make a clear proposal,
you need to have a thoughtful discussion with clear arguments, you need
to be confident that enough users we participate to the poll, and you
need to take some time and energy to actually run the poll.

But besides that, is there any reason why no poll has been made so far?

This is not a rhetoric question: maybe Emacs history has shown polls are
not effective, leading to more debate and less consensus; or maybe users
don't like polls.  

But I'd be interested to understand what prevents a poll to happen.

Thanks!

-- 
 Bastien

Re: C-d deleting region considered harmful

[Eli Zaretskii]

> From: Bastien <bzg@altern.org>
> Cc: rms@gnu.org,  eliz@gnu.org,  emacs-devel@gnu.org,  Stefan Monnier <monnier@iro.umontreal.ca>,  drew.adams@oracle.com,  miles@gnu.org
> Date: Mon, 20 Sep 2010 09:56:22 +0200
> 
> But besides that, is there any reason why no poll has been made so far?

Because we know better without any polls?

Re: C-d deleting region considered harmful

[Bastien]

Eli Zaretskii <eliz@gnu.org> writes:

>> But besides that, is there any reason why no poll has been made so far?
>
> Because we know better without any polls?

The problem is we all might have a different interpretation of "we".

-- 
 Bastien

Re: C-d deleting region considered harmful

[Eli Zaretskii]

> From: Bastien <bzg@altern.org>
> Cc: lennart.borgman@gmail.com,  rms@gnu.org,  emacs-devel@gnu.org,  monnier@iro.umontreal.ca,  drew.adams@oracle.com,  miles@gnu.org
> Date: Mon, 20 Sep 2010 13:21:49 +0200
> 
> Eli Zaretskii <eliz@gnu.org> writes:
> 
> >> But besides that, is there any reason why no poll has been made so far?
> >
> > Because we know better without any polls?
> 
> The problem is we all might have a different interpretation of "we".

I meant _my_ interpretation.

Re: C-d deleting region considered harmful

[Lennart Borgman]

On Mon, Sep 20, 2010 at 5:10 AM, Drew Adams <drew.adams@oracle.com> wrote:
>> >    Thanks.  But we're discussing here what to do with C-d:
>> >    for `delete' we already decided that it should delete an
>> >    active region (by default).  Whether you agree with this
>> >    `delete' behavior is irrelevant.
>> >
>> > It won't bother anyone for `delete' to delete the region.  Making
>> > `backspace' do so is the change that will bother lots of users.
>>
>> However a lot of other people also expect C-d to delete the visible
>> region.
>
> Huh?  Where do you get that?  Who would expect that based on other applications
> or platforms?  That is certainly not what C-d does on Windows, for instance.

Eh, sorry, thanks Drew, I meant backspace not C-d.

Re: C-d deleting region considered harmful

[David Kastrup]

Eli Zaretskii <eliz@gnu.org> writes:

>> From: Bastien <bzg@altern.org>
>> Cc: lennart.borgman@gmail.com,  rms@gnu.org,  emacs-devel@gnu.org,  monnier@iro.umontreal.ca,  drew.adams@oracle.com,  miles@gnu.org
>> Date: Mon, 20 Sep 2010 13:21:49 +0200
>> 
>> Eli Zaretskii <eliz@gnu.org> writes:
>> 
>> >> But besides that, is there any reason why no poll has been made so far?
>> >
>> > Because we know better without any polls?
>> 
>> The problem is we all might have a different interpretation of "we".
>
> I meant _my_ interpretation.

Shouldn't that read "We meant _our_ interpretation" for consistency?

-- 
David Kastrup

Re: C-d deleting region considered harmful

[Deniz Dogan]

2010/9/20 David Kastrup <dak@gnu.org>:
> Eli Zaretskii <eliz@gnu.org> writes:
>
>>> From: Bastien <bzg@altern.org>
>>> Cc: lennart.borgman@gmail.com,  rms@gnu.org,  emacs-devel@gnu.org,  monnier@iro.umontreal.ca,  drew.adams@oracle.com,  miles@gnu.org
>>> Date: Mon, 20 Sep 2010 13:21:49 +0200
>>>
>>> Eli Zaretskii <eliz@gnu.org> writes:
>>>
>>> >> But besides that, is there any reason why no poll has been made so far?
>>> >
>>> > Because we know better without any polls?
>>>
>>> The problem is we all might have a different interpretation of "we".
>>
>> I meant _my_ interpretation.
>
> Shouldn't that read "We meant _our_ interpretation" for consistency?
>

I think Eli means "his" interpretation of "we". I wonder how we should
interpret that...

-- 
Deniz Dogan

Re: C-d deleting region considered harmful

[David Kastrup]

Deniz Dogan <deniz.a.m.dogan@gmail.com> writes:

> 2010/9/20 David Kastrup <dak@gnu.org>:
>> Eli Zaretskii <eliz@gnu.org> writes:
>>
>>>> From: Bastien <bzg@altern.org>
>>>> Cc: lennart.borgman@gmail.com,  rms@gnu.org,  emacs-devel@gnu.org,  monnier@iro.umontreal.ca,  drew.adams@oracle.com,  miles@gnu.org
>>>> Date: Mon, 20 Sep 2010 13:21:49 +0200
>>>>
>>>> Eli Zaretskii <eliz@gnu.org> writes:
>>>>
>>>> >> But besides that, is there any reason why no poll has been made so far?
>>>> >
>>>> > Because we know better without any polls?
>>>>
>>>> The problem is we all might have a different interpretation of "we".
>>>
>>> I meant _my_ interpretation.
>>
>> Shouldn't that read "We meant _our_ interpretation" for consistency?
>
> I think Eli means "his" interpretation of "we".

For our interpretation of "his"?

> I wonder how we should interpret that...

I wonder why we should interpret that.

-- 
David Kastrup

Re: C-d deleting region considered harmful

[Eli Zaretskii]

> From: David Kastrup <dak@gnu.org>
> Date: Mon, 20 Sep 2010 14:07:09 +0200
> 
> Eli Zaretskii <eliz@gnu.org> writes:
> 
> >> From: Bastien <bzg@altern.org>
> >> Cc: lennart.borgman@gmail.com,  rms@gnu.org,  emacs-devel@gnu.org,  monnier@iro.umontreal.ca,  drew.adams@oracle.com,  miles@gnu.org
> >> Date: Mon, 20 Sep 2010 13:21:49 +0200
> >> 
> >> Eli Zaretskii <eliz@gnu.org> writes:
> >> 
> >> >> But besides that, is there any reason why no poll has been made so far?
> >> >
> >> > Because we know better without any polls?
> >> 
> >> The problem is we all might have a different interpretation of "we".
> >
> > I meant _my_ interpretation.
> 
> Shouldn't that read "We meant _our_ interpretation" for consistency?

That would be a circular definition, suffering from the same issue
that Bastien raised above.  But since _my_ and _ours_ is the same
here, I've elected to disambiguate it.

Re: C-d deleting region considered harmful

[Eli Zaretskii]

> From: David Kastrup <dak@gnu.org>
> Date: Mon, 20 Sep 2010 14:37:57 +0200
> 
> Deniz Dogan <deniz.a.m.dogan@gmail.com> writes:
> 
> > 2010/9/20 David Kastrup <dak@gnu.org>:
> >> Eli Zaretskii <eliz@gnu.org> writes:
> >>
> >>>> From: Bastien <bzg@altern.org>
> >>>> Cc: lennart.borgman@gmail.com,  rms@gnu.org,  emacs-devel@gnu.org,  monnier@iro.umontreal.ca,  drew.adams@oracle.com,  miles@gnu.org
> >>>> Date: Mon, 20 Sep 2010 13:21:49 +0200
> >>>>
> >>>> Eli Zaretskii <eliz@gnu.org> writes:
> >>>>
> >>>> >> But besides that, is there any reason why no poll has been made so far?
> >>>> >
> >>>> > Because we know better without any polls?
> >>>>
> >>>> The problem is we all might have a different interpretation of "we".
> >>>
> >>> I meant _my_ interpretation.
> >>
> >> Shouldn't that read "We meant _our_ interpretation" for consistency?
> >
> > I think Eli means "his" interpretation of "we".
> 
> For our interpretation of "his"?
> 
> > I wonder how we should interpret that...
> 
> I wonder why we should interpret that.

Yes.

RE: C-d deleting region considered harmful

[Drew Adams]

> If you just want to reactivate the region, the proper 
> keysequence to do only that and nothing else would be
> C-x C-x C-x C-x (otherwise you also move point).
>
> So one could make this reactivate only on second call
> (like C-SPC does).  On the other hand, C-x C-x C-x C-x is 
> often used for checking "did I leave the mark just where intended?".

Yes, I mentioned all of that.  But C-x C-x C-x could activate the region and
swap positions, and C-x C-x C-x C-x could do so without swapping positions.  Not
ideal, but a possibility.

(But it would be better to find a simple key that is not used or whose current
use could be sacrificed.)

Re: C-d deleting region considered harmful

[Miles Bader]

David Kastrup <dak@gnu.org> writes:
>> Not a good idea, I think, since one of the primary uses of C-x C-x is to
>> re-activate the region!  [for me about 90% of the time I think...]
>
> If you just want to reactivate the region, the proper keysequence to do
> only that and nothing else would be C-x C-x C-x C-x (otherwise you also
> move point).  So one could make this reactivate only on second call
> (like C-SPC does).

Er, no.

It doesn't actually _matter_ about 90% of the time whether the point and
mark are swapped (as typically the entire point of activating the region
is to operate on the region as a unit), and I think you can see that a
two-keystroke command is preferable to a four-keystroke command...

-Miles

-- 
Selfish, adj. Devoid of consideration for the selfishness of others.

Re: C-d deleting region considered harmful

[Chong Yidong]

Chong Yidong <cyd@stupidchicken.com> writes:

> Stefan Monnier <monnier@IRO.UMontreal.CA> writes:
>
>> I don't know what was Chong's motivation for it, but one reason is
>> that the `delete' key is often remapped to C-d.
>
> Indeed.  I suppose we could get away with mapping C-d to delete-char and
> [delete] to delete-forward-char.  The main disadvantage would be the
> breaking of the equivalence of C-d and DEL.

It is also worth noting that one of the motivations for making deletion
characters act on the region was to eliminate mouse-region-delete-keys,
and hence to avoid treating "regions made with the mouse" as a special
case.

[delete] was in mouse-region-delete-keys, but not C-d.  So, if C-d
deleting the active region is especially bothersome, I think it is OK to
decouple [delete] and C-d.

Re: C-d deleting region considered harmful

[Chad Brown]

On Sep 19, 2010, at 6:21 PM, Miles Bader wrote:
>> Perhaps `C-x C-x' is the prime annoyance factor here?  It certainly was
>> for me.  Perhaps it just shouldn't activate the transient mark mode?
> 
> Not a good idea, I think, since one of the primary uses of C-x C-x is to
> re-activate the region!  [for me about 90% of the time I think...]

Not to point the ``you're wrong!'' finger, but I suspect that this is not the 
primary use of C-xC-x for *most* people, but instead is the use for a
narrow selection of people who learned to appreciate using the active 
region with the keyboard before mouse/shift-selection was really common.

I suspect that the people (like me, admittedly) who started to really learn
emacs before transient regions had already internalized C-xC-x as `hop 
between two spots', and the newer people who are used to graphical
feedback flourishes like a highlighted active region are used to setting the
mark explicitly, or using either the mouse or shift-selection.

I suggest adding a flag to invert the default meaning of ARG in
exchange-point-and-mark, at least as an experiment.  I suspect that
the new behavior would be more in line with expected behavior for most
people, but it'd be hard to determine, and it seems reasonable to
default to the existing behavior in such a situation.

I'll see if I can put together a patch and try it out.

*Chad

Re: C-d deleting region considered harmful

[Richard Stallman]

    > It won't bother anyone for `delete' to delete the region.  Making
    > `backspace' do so is the change that will bother lots of users.

    But it's also the only such change that really matters for
    newbie-friendless.

I am not convinced it does.  The case in which newbies would expect
Backspace to delete the region is after mouse-selecting it,
and it already does delete the region in that case.

Re: C-d deleting region considered harmful

[Lennart Borgman]

On Tue, Sep 21, 2010 at 2:56 AM, Richard Stallman <rms@gnu.org> wrote:
>    > It won't bother anyone for `delete' to delete the region.  Making
>    > `backspace' do so is the change that will bother lots of users.
>
>    But it's also the only such change that really matters for
>    newbie-friendless.
>
> I am not convinced it does.  The case in which newbies would expect
> Backspace to delete the region is after mouse-selecting it,
> and it already does delete the region in that case.


I think a newbie expect the visible region to be delete by backspace
always (at least I think that is the case on w32). Why don't you think
that?

Re: C-d deleting region considered harmful

[Miles Bader]

Chad Brown <yandros@MIT.EDU> writes:
>> Not a good idea, I think, since one of the primary uses of C-x C-x
>> is to re-activate the region!  [for me about 90% of the time I
>> think...]
>
> Not to point the ``you're wrong!'' finger, but I suspect that this
> is not the primary use of C-xC-x for *most* people, but instead is
> the use for a narrow selection of people who learned to appreciate
> using the active region with the keyboard before
> mouse/shift-selection was really common.

It's true, that my claim is "just my claim" -- but on the other hand
so is yours (it doesn't seem any more logically consistent).  I think
actual data is necessary to actually determine which is really more
widespread.

> I suspect that the people (like me, admittedly) who started to
> really learn emacs before transient regions had already internalized
> C-xC-x as `hop between two spots', and the newer people who are used
> to graphical feedback flourishes like a highlighted active region
> are used to setting the mark explicitly, or using either the mouse
> or shift-selection.

Remember, C-x C-x _reactivates_ the region, it's not a "set the region
command", and indeed, my most common use seems to be using it after
some typing goof accidentally causes my (laboriously selected) region
to get deactivated for some reason.

For the record, I'm definitely in your generation -- I first started
using Emacs regularly in 1983.  So while I like tmm a lot, I'm of the
"keyboard generation."  I learned the "new usage" of C-x C-x pretty
quickly, and I don't see any obvious reason to believe that others
long-time users are any different (many long-time users of course,
hate tmm generally, but they probably just turn it off, and so really
don't have a dog in this race).

> I suggest adding a flag to invert the default meaning of ARG in
> exchange-point-and-mark, at least as an experiment.  I suspect that
> the new behavior would be more in line with expected behavior for most
> people, but it'd be hard to determine, and it seems reasonable to
> default to the existing behavior in such a situation.

Er, how about a bit more basis than "I suspect" first...

-Miles

-- 
Abstainer, n. A weak person who yields to the temptation of denying himself a
pleasure. A total abstainer is one who abstains from everything but
abstention, and especially from inactivity in the affairs of others.

Re: C-d deleting region considered harmful

[Miles Bader]

Richard Stallman <rms@gnu.org> writes:

>     But it's also the only such change that really matters for
>     newbie-friendless.
>
> I am not convinced it does.  The case in which newbies would expect
> Backspace to delete the region is after mouse-selecting it,

No.  Newbies actually use the keyboard too, you know (even if they're
not used to Emacs)...

Anyway, having mouse-regions be "magic" is confusingly inconsistent, and
a bad UI generally.  That was a big wart on Emacs in the past.

It's obviously impossible to have perfect defaults, because we're trying
to unify a widely disparate set of user experiences into one UI.  A
change that makes things better for some often makes others unhappy.

There's little we can do other than trying to minimize unhappiness, but
at least we can (1) keep things customizable, so those that just can't
deal with a change can easily find relief, and (2) try to make Emacs' UI
broadly consistent, so that's it's at least easy for users to form a
mental-model of how things work (and easy to document)...

-Miles

-- 
Carefully crafted initial estimates reward you not only with
reduced computational effort, but also with understanding and
increased self-esteem.         -- Numerical methods in C,
  Chapter 9. "Root Finding and Nonlinear Sets of Equations"

Re: C-d deleting region considered harmful

[Leo]

On 2010-09-21 02:30 +0100, Miles Bader wrote:
> There's little we can do other than trying to minimize unhappiness,
> but at least we can (1) keep things customizable, so those that just
> can't deal with a change can easily find relief, and (2) try to make
> Emacs' UI broadly consistent, so that's it's at least easy for users
> to form a mental-model of how things work (and easy to document)...

+1 for consistency.

BTW, I think the "Newbie Argument" start to smell bad. It merely leads
to bad decisions. Sometimes I don't even know if people are talking
about the same kind of newbies.

Leo

Re: C-d deleting region considered harmful

[Miles Bader]

Leo <sdl.web@gmail.com> writes:
> BTW, I think the "Newbie Argument" start to smell bad. It merely leads
> to bad decisions. Sometimes I don't even know if people are talking
> about the same kind of newbies.

You can ignore the "newbie argument" if you want; I think a lot of
what's being done in their name actually applies equally well to
people that switch often between Emacs and other UIs, and generally to
the goal of having Emacs exist harmoniously in today's environments.
Maybe some people _only_ use Emacs, but I think many of us aren't so
lucky...  :)

For instance, I find having shift-selection available in emacs very
nice for those times when I happen to be flipping back-and-forth
between Emacs and another text-editing interface.  Of course,
shift-select doesn't interfere with other features, and other less
benign changes need to be approached more carefully.

[I turn on "emacs bindings" modes whenever I can (e.g. GTK and
Eclipse), but these are usually only very approximate emulations, and
one still needs to use the native bindings for many operations.]

-Miles

-- 
Bigot, n. One who is obstinately and zealously attached to an opinion that
you do not entertain.

Re: C-d deleting region considered harmful

[Stephen J. Turnbull]

Christoph writes:
 > On 9/18/2010 3:18 PM, Eli Zaretskii wrote:
 > 
 > >> So I just switched off `transient-mark-mode', which is something
 > >> I suspect most Emacs old-timers will be more comfortable with.
 > >
 > > Here's one such old-timer.
 > Out of curiosity, from old-timer to new-timer, what advantages does
 > the traditional Emacs behavior over transient-mark-mode?

Mostly, it's traditional and old-timers are used to it.  It took me
close to a month to get used to the various differences, then I
decided I liked t-m-m (actually, zmacs-regions) better on than off.

More scientifically, if you have an active region, then you can have
"modal" behavior: deletion operations can (implicitly) act on the
region instead of on specific text units, insertion operations can
(implicitly) substitute new text for the region, and so on.  Without
active regions, you can't have this kind of modal behavior.

The two styles are *equally* powerful.  Some people like the modal,
DWIMmish, behavior better (it can be slightly more efficient in terms
of keystroke count), while others like the non-modal, DWIS ("do what I
say"), behavior better (it's better adapted to creating personal
idioms and using "muscle memory", I think).  Much of the taste
difference can be attributed to "what you are used to", of course, and
I think that the strongest reasons for preferring one to the other are
what you are used to as "traditional" for you.

Re: C-d deleting region considered harmful

[David Kastrup]

Lennart Borgman <lennart.borgman@gmail.com> writes:

> On Tue, Sep 21, 2010 at 2:56 AM, Richard Stallman <rms@gnu.org> wrote:
>>    > It won't bother anyone for `delete' to delete the region.  Making
>>    > `backspace' do so is the change that will bother lots of users.
>>
>>    But it's also the only such change that really matters for
>>    newbie-friendless.
>>
>> I am not convinced it does.  The case in which newbies would expect
>> Backspace to delete the region is after mouse-selecting it,
>> and it already does delete the region in that case.
>
>
> I think a newbie expect the visible region to be delete by backspace
> always (at least I think that is the case on w32). Why don't you think
> that?

Is there _any_ way to create an active region under w32 that does not
already do delete-by-backspace in Emacs?

-- 
David Kastrup

Re: C-d deleting region considered harmful

[David Kastrup]

"Stephen J. Turnbull" <stephen@xemacs.org> writes:

> Christoph writes:
>  > On 9/18/2010 3:18 PM, Eli Zaretskii wrote:
>  > 
>  > >> So I just switched off `transient-mark-mode', which is something
>  > >> I suspect most Emacs old-timers will be more comfortable with.
>  > >
>  > > Here's one such old-timer.
>  > Out of curiosity, from old-timer to new-timer, what advantages does
>  > the traditional Emacs behavior over transient-mark-mode?
>
> Mostly, it's traditional and old-timers are used to it.  It took me
> close to a month to get used to the various differences, then I
> decided I liked t-m-m (actually, zmacs-regions) better on than off.

That is to be taken with a heavy grain of salt since historically
zmacs-regions and transient-mark-mode have had a number of small
differences.  Enough that some people moving back and forth used
zmacs-regions on XEmacs, but scorned transient-mark-mode on Emacs as
being unusable.

Now the latter has been in constant flux over the last versions.  I
don't know how they compare in user acceptance and semantics in the
current state.

And I don't know whether there are any Double Power Users left who could
give qualified comparisons.

> The two styles are *equally* powerful.  Some people like the modal,
> DWIMmish, behavior better (it can be slightly more efficient in terms
> of keystroke count), while others like the non-modal, DWIS ("do what I
> say"), behavior better (it's better adapted to creating personal
> idioms and using "muscle memory", I think).  Much of the taste
> difference can be attributed to "what you are used to", of course, and
> I think that the strongest reasons for preferring one to the other are
> what you are used to as "traditional" for you.

I don't think that this really applies all too much for
transient-mark-mode: the complaints do not center around user interface
philosophies, but rather the nitty gritty details.

In short: figuring out for each choice the least annoying details of
operation.

Not having transient regions as sideeffects of other useful operations
at all, of course, is reasonably simple to implement.

There are not really many commands left where the setting of
transient-mark-mode should make a difference, or even a bad surprise.  I
can think of C-SPC, C-x C-x, M-< and M-> and that's more or less it.
Then there are the mark-something commands (I have my doubts they are
used very much) where the transient-region behavior would seem somewhat
less contentious.

-- 
David Kastrup

Re: C-d deleting region considered harmful

[PJ Weisberg]

On Tue, Sep 21, 2010 at 12:51 AM, David Kastrup <dak@gnu.org> wrote:

> Is there _any_ way to create an active region under w32 that does not
> already do delete-by-backspace in Emacs?

Holding shift while moving your cursor over a region with the arrow keys.

Re: C-d deleting region considered harmful

[Leo]

On 2010-09-21 07:29 +0100, Stephen J. Turnbull wrote:
> More scientifically, if you have an active region, then you can have
> "modal" behavior: deletion operations can (implicitly) act on the
> region instead of on specific text units, insertion operations can
> (implicitly) substitute new text for the region, and so on. Without
> active regions, you can't have this kind of modal behavior.

Yeah, that is useful. Some prime keys can be freed for better things,
for example, C-x C-l and C-x C-u.

Leo

Re: C-d deleting region considered harmful

[Deniz Dogan]

2010/9/21 Leo <sdl.web@gmail.com>:
> On 2010-09-21 07:29 +0100, Stephen J. Turnbull wrote:
>> More scientifically, if you have an active region, then you can have
>> "modal" behavior: deletion operations can (implicitly) act on the
>> region instead of on specific text units, insertion operations can
>> (implicitly) substitute new text for the region, and so on. Without
>> active regions, you can't have this kind of modal behavior.
>
> Yeah, that is useful. Some prime keys can be freed for better things,
> for example, C-x C-l and C-x C-u.
>

I'm not sure what you mean. I use C-x C-l and C-x C-u on a nearly
daily basis. If M-l and M-u executed downcase-region and upcase-region
if the region was active, I'd be happy to use that instead.

-- 
Deniz Dogan

Re: C-d deleting region considered harmful

[Leo]

On 2010-09-21 09:52 +0100, Deniz Dogan wrote:
> I'm not sure what you mean. I use C-x C-l and C-x C-u on a nearly
> daily basis. If M-l and M-u executed downcase-region and upcase-region
> if the region was active, I'd be happy to use that instead.

Yes, I rebind M-l/u to something that does that and bind C-x C-l/u to
something different.

Leo

Re: C-d deleting region considered harmful

[Stephen J. Turnbull]

Deniz Dogan writes:

 > I'm not sure what you mean. I use C-x C-l and C-x C-u on a nearly
 > daily basis. If M-l and M-u executed downcase-region and
 > upcase-region if the region was active, I'd be happy to use that
 > instead.

Indeed, that's what XEmacs does.  I personally find it very useful to
have that DWIM behavior, and have never heard a complaint about it.

Re: C-d deleting region considered harmful

[Chad Brown]

On Sep 20, 2010, at 6:03 PM, Miles Bader wrote:
> It's true, that my claim is "just my claim" -- but on the other hand
> so is yours (it doesn't seem any more logically consistent). 

True enough.  My opinions on `expected behavior' are perhaps better
described ``what I believe new users who start with modern GUI's
(Gnome/KDE/MacOSX/Winwhatever) are more likely to expect, based on a
pretty small sample size'', which is indeed `suspect' by itself.

>> I suggest adding a flag to invert the default meaning of ARG in
>> exchange-point-and-mark, at least as an experiment.  I suspect that
>> the new behavior would be more in line with expected behavior for most
>> people, but it'd be hard to determine, and it seems reasonable to
>> default to the existing behavior in such a situation.
> 
> Er, how about a bit more basis than "I suspect" first...

Probably I should have emphasized the ``as an experiment'' part more.
That's my fault; I had originally intended to include a patch to try, then
something came up that caused me to not have as much time as I'd
expected so I quickly rewrote the email and sent it.  I wasn't suggesting
that we change the default for everyone and see what happens, but
it certainly could be read that way.

I'm using such a simple patch now, but I don't believe that it will do
the right thing in the face of t-m-m being enabled/disabled, so I'm
holding out for more time or for someone else to create and send a 
better experimental candidate.  At that point, I'd hope to get some of 
the list stalwarts who dislike t-m-m to potentially give it a try and see
if it's better, worse, or indistinct.   

Sorry for any confusion.
*Chad

Re: C-d deleting region considered harmful

[Richard Stallman]

    This is not a rhetoric question: maybe Emacs history has shown polls are
    not effective, leading to more debate and less consensus; or maybe users
    don't like polls.  

I did on the order of 10 polls while I was Emacs maintainer,
and I am very glad I did.  They gave a lot of useful information.

Re: C-d deleting region considered harmful

[Miles Bader]

Richard Stallman <rms@gnu.org> writes:
>     This is not a rhetoric question: maybe Emacs history has shown polls are
>     not effective, leading to more debate and less consensus; or maybe users
>     don't like polls.  
>
> I did on the order of 10 polls while I was Emacs maintainer,
> and I am very glad I did.  They gave a lot of useful information.

[I seem to recall you making the point in the past that the polls are
not so much a "vote" as a way to solicit reasoned arguments from a
somewhat wider group than the regular developers.]

Of course, the Emacs user community does change, and probably is more
disparate now than it was in the past (with more people getting Emacs,
and Emacs support, indirectly through N levels, rather than directly
using emacs-help etc).

It seems like the usefulness of the results would depend very much on
the nature of the question -- e.g., the results might be a lot better
when you want the feedback of experienced long-time Emacs users (and
it's OK for the results to be heavily biased in that direction), than
when you want the opinion of the "mythical newbie" or even the average
Emacs user.

-miles

-- 
It wasn't the Exxon Valdez captain's driving that caused the Alaskan oil spill.
It was yours.  [Greenpeace advertisement, New York Times, 25 February 1990]

Re: C-d deleting region considered harmful

[Eli Zaretskii]

> Date: Sun, 19 Sep 2010 21:58:07 +0200
> From: Eli Zaretskii <eliz@gnu.org>
> Cc: emacs-devel@gnu.org
> 
> > From: David Kastrup <dak@gnu.org>
> > Date: Sun, 19 Sep 2010 21:18:12 +0200
> > 
> > I just did C-h h C-x C-q, went into the Hebrew text and typed a few
> > characters (like x or ( ) since ( ) are not strictly L-R).
> > 
> > I got seasick.
> 
> I hope your keyboard or input method mirror characters, or else you
> will really become seasick.

Btw, I wrote the entire Hebrew translation of the tutorial with Emacs,
using the native Windows Hebrew keyboard, and found the existing Emacs
facilities quite adequate for writing even such large and complex
document in a R2L script.

Re: C-d deleting region considered harmful

[Richard Stallman]

    [I seem to recall you making the point in the past that the polls are
    not so much a "vote" as a way to solicit reasoned arguments from a
    somewhat wider group than the regular developers.]

Yes, indeed.

    It seems like the usefulness of the results would depend very much on
    the nature of the question -- e.g., the results might be a lot better
    when you want the feedback of experienced long-time Emacs users (and
    it's OK for the results to be heavily biased in that direction), than
    when you want the opinion of the "mythical newbie" or even the average
    Emacs user.

What we find out from polls is whether significant groups of people
find a certain feature useful, and if so _why_ they do.
That enables us to find changes that won't cause problems for them
and avoid changes that will.

If we just speculate about this, we might be very wrong.