"C-x 5 5" vs "C-x 4 1" inconsistency

"C-x 5 5" vs "C-x 4 1" inconsistency

[Eli Zaretskii]

I think this inconsistency is unfortunate and unjustified.  Can we
please move same-window-prefix to "C-x 4 5" for consistency with the
equivalent frame command?

Re: "C-x 5 5" vs "C-x 4 1" inconsistency

[Eli Zaretskii]

> Date: Sat, 09 Oct 2021 12:26:09 +0300
> From: Eli Zaretskii <eliz@gnu.org>
> 
> I think this inconsistency is unfortunate and unjustified.  Can we
> please move same-window-prefix to "C-x 4 5" for consistency with the
> equivalent frame command?

Sorry, I meant "C-x 4 4" and other-window-prefix.

Re: "C-x 5 5" vs "C-x 4 1" inconsistency

[Juri Linkov]

>> I think this inconsistency is unfortunate and unjustified.  Can we
>> please move same-window-prefix to "C-x 4 5" for consistency with the
>> equivalent frame command?
>
> Sorry, I meant "C-x 4 4" and other-window-prefix.

Please explain where do you think is inconsistency.
'C-x 4 4' is already bound to 'other-window-prefix'.

Re: "C-x 5 5" vs "C-x 4 1" inconsistency

[Kévin Le Gouguec]

Juri Linkov <juri@linkov.net> writes:

>>> I think this inconsistency is unfortunate and unjustified.  Can we
>>> please move same-window-prefix to "C-x 4 5" for consistency with the
>>> equivalent frame command?
>>
>> Sorry, I meant "C-x 4 4" and other-window-prefix.
>
> Please explain where do you think is inconsistency.
> 'C-x 4 4' is already bound to 'other-window-prefix'.

IIUC Eli would find C-x 4 5 more consistent for other-window-prefix (I
hope I ran the substitutions correctly; apologies if not).

Personally I like the current arrangement: I think of it as "C-x X X for
other-Y-prefix", with

- X=4 ⇒ Y=window
- X=5 ⇒ Y=frame
- X=t ⇒ Y=tab

I find "mash X to run in new Y" simpler to remember than "press X then 5
to run in new Y".

Re: "C-x 5 5" vs "C-x 4 1" inconsistency

[Stefan Kangas]

Kévin Le Gouguec <kevin.legouguec@gmail.com> writes:

> I find "mash X to run in new Y" simpler to remember than "press X then 5
> to run in new Y".

I very much agree, the new binding is logical and easy to remember to me:

C-x 4 4    other-window-prefix
C-x 5 5    other-frame-prefix

Re: "C-x 5 5" vs "C-x 4 1" inconsistency

[Eli Zaretskii]

> From: Juri Linkov <juri@linkov.net>
> Cc: emacs-devel@gnu.org
> Date: Sun, 10 Oct 2021 11:10:50 +0300
> 
> >> I think this inconsistency is unfortunate and unjustified.  Can we
> >> please move same-window-prefix to "C-x 4 5" for consistency with the
> >> equivalent frame command?
> >
> > Sorry, I meant "C-x 4 4" and other-window-prefix.
> 
> Please explain where do you think is inconsistency.
> 'C-x 4 4' is already bound to 'other-window-prefix'.

But other-frame-prefix is "C-x 5 5".  This is inconsistent with other
uses of these prefixes, specifically "C-x 5 2", "C-x 5 b", C-x 5 f",
which have their "C-x 4" counterparts.

Re: "C-x 5 5" vs "C-x 4 1" inconsistency

[Daniel Martín]

Eli Zaretskii <eliz@gnu.org> writes:

>> From: Juri Linkov <juri@linkov.net>
>> Cc: emacs-devel@gnu.org
>> Date: Sun, 10 Oct 2021 11:10:50 +0300
>> 
>> >> I think this inconsistency is unfortunate and unjustified.  Can we
>> >> please move same-window-prefix to "C-x 4 5" for consistency with the
>> >> equivalent frame command?
>> >
>> > Sorry, I meant "C-x 4 4" and other-window-prefix.
>> 
>> Please explain where do you think is inconsistency.
>> 'C-x 4 4' is already bound to 'other-window-prefix'.
>
> But other-frame-prefix is "C-x 5 5".  This is inconsistent with other
> uses of these prefixes, specifically "C-x 5 2", "C-x 5 b", C-x 5 f",
> which have their "C-x 4" counterparts.

I agree it'd be nice to be consistent.  The argument about "C-x 4 4" and
"C-x 5 5" being easier to remember might not be relevant once the
commands are in muscle memory, I don't know.  I don't have a strong
opinion here.

Re: "C-x 5 5" vs "C-x 4 1" inconsistency

[Juri Linkov]

>> >> I think this inconsistency is unfortunate and unjustified.  Can we
>> >> please move same-window-prefix to "C-x 4 5" for consistency with the
>> >> equivalent frame command?
>> >
>> > Sorry, I meant "C-x 4 4" and other-window-prefix.
>>
>> Please explain where do you think is inconsistency.
>> 'C-x 4 4' is already bound to 'other-window-prefix'.
>
> But other-frame-prefix is "C-x 5 5".  This is inconsistent with other
> uses of these prefixes, specifically "C-x 5 2", "C-x 5 b", C-x 5 f",
> which have their "C-x 4" counterparts.

The prefix 'C-x 5' is for frames, so 'C-x 5 5' creates a new frame.
The prefix 'C-x 4' is for windows, so 'C-x 4 4' creates a new window.
Also like 'C-x 1' keeps the selected window, 'C-x 4 1' shows the next
buffer in the same window while keeping it selected.

Recently Drew mentioned that there is the command 'tear-off-window'.
It would perfectly fit into the keybinding 'C-x 4 5' with mnemonics
that '4' (window) is detached to '5' (frame):

diff --git a/lisp/window.el b/lisp/window.el
index 971264b634..2fd9c970b0 100644
--- a/lisp/window.el
+++ b/lisp/window.el
@@ -10381,6 +10381,7 @@ ctl-x-map
 (define-key ctl-x-4-map "0" 'kill-buffer-and-window)
 (define-key ctl-x-4-map "1" 'same-window-prefix)
 (define-key ctl-x-4-map "4" 'other-window-prefix)
+(define-key ctl-x-4-map "5" 'tear-off-window)
 
 (defvar other-window-repeat-map
   (let ((map (make-sparse-keymap)))
-- 

Re: "C-x 5 5" vs "C-x 4 1" inconsistency

[Eli Zaretskii]

> From: Juri Linkov <juri@linkov.net>
> Cc: emacs-devel@gnu.org
> Date: Sun, 10 Oct 2021 20:15:14 +0300
> 
> >> >> I think this inconsistency is unfortunate and unjustified.  Can we
> >> >> please move same-window-prefix to "C-x 4 5" for consistency with the
> >> >> equivalent frame command?
> >> >
> >> > Sorry, I meant "C-x 4 4" and other-window-prefix.
> >>
> >> Please explain where do you think is inconsistency.
> >> 'C-x 4 4' is already bound to 'other-window-prefix'.
> >
> > But other-frame-prefix is "C-x 5 5".  This is inconsistent with other
> > uses of these prefixes, specifically "C-x 5 2", "C-x 5 b", C-x 5 f",
> > which have their "C-x 4" counterparts.
> 
> The prefix 'C-x 5' is for frames, so 'C-x 5 5' creates a new frame.
> The prefix 'C-x 4' is for windows, so 'C-x 4 4' creates a new window.
> Also like 'C-x 1' keeps the selected window, 'C-x 4 1' shows the next
> buffer in the same window while keeping it selected.

And what do you say about the other examples I gave above?

It was always the rule with these commands that similar operations
have the same "final" keys.  Now you've broken this.  I think this is
not a good situation.  It will definitely get in the way of my
remembering which one is which.

> Recently Drew mentioned that there is the command 'tear-off-window'.
> It would perfectly fit into the keybinding 'C-x 4 5' with mnemonics
> that '4' (window) is detached to '5' (frame):

I don't see how this weak mnemonics would help with the broken
mnemonics of "C-x 4 4" vs "C-x 5 5".  The prefixes should start with
"C-x 4/5" and continue with the same character.  If you don't want to
use 4 or 5 for both, let's find a different letter, but make it
identical for both "C-x 4" and "C-x 5".

Re: "C-x 5 5" vs "C-x 4 1" inconsistency

[Juri Linkov]

>> The prefix 'C-x 5' is for frames, so 'C-x 5 5' creates a new frame.
>> The prefix 'C-x 4' is for windows, so 'C-x 4 4' creates a new window.
>> Also like 'C-x 1' keeps the selected window, 'C-x 4 1' shows the next
>> buffer in the same window while keeping it selected.
>
> And what do you say about the other examples I gave above?
>
> It was always the rule with these commands that similar operations
> have the same "final" keys.  Now you've broken this.  I think this is
> not a good situation.  It will definitely get in the way of my
> remembering which one is which.

All current keybindings already have the same final keys, e.g.

C-x 4 4 C-h i   - opens Info in a new window
C-x 5 5 C-h i   - opens Info in a new frame

Their final keys are the same.

Re: "C-x 5 5" vs "C-x 4 1" inconsistency

[Eli Zaretskii]

> From: Juri Linkov <juri@linkov.net>
> Date: Sun, 10 Oct 2021 22:02:43 +0300
> Cc: emacs-devel@gnu.org
> 
> >> The prefix 'C-x 5' is for frames, so 'C-x 5 5' creates a new frame.
> >> The prefix 'C-x 4' is for windows, so 'C-x 4 4' creates a new window.
> >> Also like 'C-x 1' keeps the selected window, 'C-x 4 1' shows the next
> >> buffer in the same window while keeping it selected.
> >
> > And what do you say about the other examples I gave above?
> >
> > It was always the rule with these commands that similar operations
> > have the same "final" keys.  Now you've broken this.  I think this is
> > not a good situation.  It will definitely get in the way of my
> > remembering which one is which.
> 
> All current keybindings already have the same final keys, e.g.
> 
> C-x 4 4 C-h i   - opens Info in a new window
> C-x 5 5 C-h i   - opens Info in a new frame
> 
> Their final keys are the same.

That's not what I meant, and you know it.

Re: "C-x 5 5" vs "C-x 4 1" inconsistency

[Juri Linkov]

>> All current keybindings already have the same final keys, e.g.
>>
>> C-x 4 4 C-h i   - opens Info in a new window
>> C-x 5 5 C-h i   - opens Info in a new frame
>>
>> Their final keys are the same.
>
> That's not what I meant, and you know it.

What I definitely know that here is some
misunderstanding that needs clarification.
The Info node (info "(emacs) Keys") says:

  ‘C-x’ and ‘C-x 4’ are prefix keys.

The same way 'C-x 4 4' is a prefix key as well.

There is no reason to change this prefix key
to something more hard to type such as 'C-x 4 x y z ...'

Re: "C-x 5 5" vs "C-x 4 1" inconsistency

[Stefan Kangas]

Eli Zaretskii <eliz@gnu.org> writes:

>> The prefix 'C-x 5' is for frames, so 'C-x 5 5' creates a new frame.
>> The prefix 'C-x 4' is for windows, so 'C-x 4 4' creates a new window.
>> Also like 'C-x 1' keeps the selected window, 'C-x 4 1' shows the next
>> buffer in the same window while keeping it selected.
>
> And what do you say about the other examples I gave above?
>
> It was always the rule with these commands that similar operations
> have the same "final" keys.  Now you've broken this.  I think this is
> not a good situation.  It will definitely get in the way of my
> remembering which one is which.

I don't see that we need to be consistent in the sense you propose.
We could instead be consistent in the sense we are now.

It feels quite natural to say `C-x 5 5', because here the final "5"
refers to another frame.  Whereas in `C-x 4 4', the final "4" refers to
another window.  Highly logical and consistent to me.

`C-x 4 4' is also more ergonomic.  I appreciate that, since I tend to
use this command quite a bit.

Re: "C-x 5 5" vs "C-x 4 1" inconsistency

[Eli Zaretskii]

> Date: Sun, 10 Oct 2021 21:07:16 +0300
> From: Eli Zaretskii <eliz@gnu.org>
> Cc: emacs-devel@gnu.org
> 
> If you don't want to use 4 or 5 for both, let's find a different
> letter, but make it identical for both "C-x 4" and "C-x 5".

How about "C-x 4 x" and "C-x 5 x". where "x" stands for "eXecute"?
An alternative could be "C-x 4 RET" and "C-x 5 RET".

Would any of these be acceptable?

Re: "C-x 5 5" vs "C-x 4 1" inconsistency

[Kévin Le Gouguec]

Eli Zaretskii <eliz@gnu.org> writes:

>> Date: Sun, 10 Oct 2021 21:07:16 +0300
>> From: Eli Zaretskii <eliz@gnu.org>
>> Cc: emacs-devel@gnu.org
>> 
>> If you don't want to use 4 or 5 for both, let's find a different
>> letter, but make it identical for both "C-x 4" and "C-x 5".
>
> How about "C-x 4 x" and "C-x 5 x". where "x" stands for "eXecute"?
> An alternative could be "C-x 4 RET" and "C-x 5 RET".
>
> Would any of these be acceptable?

I don't mind x or RET in terms of mnemonics, but my clumsy fingers find
them less ergonomic than C-x K K.

FWIW, I don't find it shocking that other-{frame,tab,window}-prefix do
not end with identical letters.  True, they are the only commands in the
{ctl-x-4,ctl-x-5,tab-prefix}-map keymaps that break this convention, but
they still follow a consistent pattern: C-x K K, where K is the prefix
key for the relevant UI element.

I also support Juri's suggestion to bind tear-off-window to C-x 4 5: I
feel like the "_4_ to _5_" sequence translates quite well to "move this
_window_ to another _frame_", at least in my head.

Re: "C-x 5 5" vs "C-x 4 1" inconsistency

[Eli Zaretskii]

> From: Kévin Le Gouguec <kevin.legouguec@gmail.com>
> Cc: juri@linkov.net,  emacs-devel@gnu.org
> Date: Mon, 11 Oct 2021 07:48:25 +0200
> 
> > How about "C-x 4 x" and "C-x 5 x". where "x" stands for "eXecute"?
> > An alternative could be "C-x 4 RET" and "C-x 5 RET".
> >
> > Would any of these be acceptable?
> 
> I don't mind x or RET in terms of mnemonics, but my clumsy fingers find
> them less ergonomic than C-x K K.

That's a peculiar argument.  There's only one "C-x K K" combination
for any K, and yet we have more than a dozen commands starting with
"C-x 4" and similarly for "C-x 5".  Does it mean that all but one of
them is not ergonomic?

Re: "C-x 5 5" vs "C-x 4 1" inconsistency

[Kévin Le Gouguec]

Eli Zaretskii <eliz@gnu.org> writes:

> That's a peculiar argument.  There's only one "C-x K K" combination
> for any K, and yet we have more than a dozen commands starting with
> "C-x 4" and similarly for "C-x 5".  Does it mean that all but one of
> them is not ergonomic?

All I meant was that given a choice between C-x K K and C-x K x, the
former feels more "ergonomic" (by which I mostly mean "durr mashing same
key twice is easy, me likes it 🤪"; dunno if Stefan's definition of the
word is more sophisticated[1]).

I did not mean to comment on the dozen other commands in
{ctl-x-4,ctl-x-5,tab-prefix}-map (I don't find C-x [45] very ergonomic
anyway on AZERTY, since digits require holding Shift).

I acknowledge that the ergonomics argument, by itself, does not justify
breaking the current convention ("similar commands end with identical
keys"), and there is no reason a priori that C-x K K, "ergonomic" as it
might be, should be given to other-X-prefix.

All I can say is that it makes some intuitive sense to me: commands
starting with C-x K ∀K∈{4,5,t} generally mean "Do/find/visit something
in another X", therefore "hit K twice to run next command in another X"
does not sound too outlandish.


[1] <CADwFkmnoMJeUGfPS5jAbYso0fWCLOrYnWuW7FcFvnZx0mpuqYg@mail.gmail.com>
    https://lists.gnu.org/archive/html/emacs-devel/2021-10/msg00765.html