Suggestion: Ace-window in Emacs core for switching windows?

Suggestion: Ace-window in Emacs core for switching windows?

[Jeremy Bryant via Emacs development discussions.]

Ace-window is a quick way to switch windows, optionally using the home
key row.

It has a high rating of 89% in ELPA
https://elpa.gnu.org/packages/

I have found it useful and also stable - It doesn't appear to undergo
major changes or need much ongoing maintenance.

Could it be a candidate for absorption into Emacs core?

Re: Suggestion: Ace-window in Emacs core for switching windows?

[Eli Zaretskii]

> Date: Tue, 25 Jun 2024 22:37:58 +0100
> From:  Jeremy Bryant via "Emacs development discussions." <emacs-devel@gnu.org>
> 
> 
> Ace-window is a quick way to switch windows, optionally using the home
> key row.
> 
> It has a high rating of 89% in ELPA
> https://elpa.gnu.org/packages/
> 
> I have found it useful and also stable - It doesn't appear to undergo
> major changes or need much ongoing maintenance.
> 
> Could it be a candidate for absorption into Emacs core?

It could be, but why is it a problem to leave it on ELPA?

Re: Suggestion: Ace-window in Emacs core for switching windows?

[Jeremy Bryant]

Eli Zaretskii <eliz@gnu.org> writes:

>> Date: Tue, 25 Jun 2024 22:37:58 +0100
>> From:  Jeremy Bryant via "Emacs development discussions." <emacs-devel@gnu.org>
>> 
>> 
>> Ace-window is a quick way to switch windows, optionally using the home
>> key row.
>> 
>> It has a high rating of 89% in ELPA
>> https://elpa.gnu.org/packages/
>> 
>> I have found it useful and also stable - It doesn't appear to undergo
>> major changes or need much ongoing maintenance.
>> 
>> Could it be a candidate for absorption into Emacs core?
>
> It could be, but why is it a problem to leave it on ELPA?

There are deployment scenarios where users can only use the Emacs
provided in the distro, with no internet access for policy or security reasons.
In particular, that means only built-in packages, without access to the internet.

As I personally use ace-window, I found it lacking in that scenario.

Re: Suggestion: Ace-window in Emacs core for switching windows?

[Philip Kaludercic]

Jeremy Bryant <jb@jeremybryant.net> writes:

> Eli Zaretskii <eliz@gnu.org> writes:
>
>>> Date: Tue, 25 Jun 2024 22:37:58 +0100
>>> From:  Jeremy Bryant via "Emacs development discussions." <emacs-devel@gnu.org>
>>> 
>>> 
>>> Ace-window is a quick way to switch windows, optionally using the home
>>> key row.
>>> 
>>> It has a high rating of 89% in ELPA
>>> https://elpa.gnu.org/packages/
>>> 
>>> I have found it useful and also stable - It doesn't appear to undergo
>>> major changes or need much ongoing maintenance.
>>> 
>>> Could it be a candidate for absorption into Emacs core?
>>
>> It could be, but why is it a problem to leave it on ELPA?
>
> There are deployment scenarios where users can only use the Emacs
> provided in the distro, with no internet access for policy or security reasons.
> In particular, that means only built-in packages, without access to the internet.
>
> As I personally use ace-window, I found it lacking in that scenario.

Given that we already have windmove, would the addition of ace-window
really change that much?

-- 
	Philip Kaludercic on peregrine

Re: Suggestion: Ace-window in Emacs core for switching windows?

[Stefan Kangas]

Philip Kaludercic <philipk@posteo.net> writes:

> Given that we already have windmove, would the addition of ace-window
> really change that much?

The same argument could be used for many packages that we have in core.
Who needs Gnus when you have rmail, and so on.  The packages do
basically the same thing, but in very different ways.

I don't use ace-window myself, but note that it is the 28th most popular
package on MELPA with 1,773,799 downloads, only slightly behind
'which-key' in popularity that we just added to core.

See: https://melpa.org/#/?sort=downloads&asc=false

Until we figure out how to bundle GNU ELPA packages, something that has
seen little progress in a long time, I think there is a case to be made
for adding some of the more popular packages to core.

Re: Suggestion: Ace-window in Emacs core for switching windows?

[Jeremy Bryant]

Stefan Kangas <stefankangas@gmail.com> writes:

> Philip Kaludercic <philipk@posteo.net> writes:
>
>> Given that we already have windmove, would the addition of ace-window
>> really change that much?
>
> The same argument could be used for many packages that we have in core.
> Who needs Gnus when you have rmail, and so on.  The packages do
> basically the same thing, but in very different ways.
>
> I don't use ace-window myself, but note that it is the 28th most popular
> package on MELPA with 1,773,799 downloads, only slightly behind
> 'which-key' in popularity that we just added to core.
>
> See: https://melpa.org/#/?sort=downloads&asc=false
>
> Until we figure out how to bundle GNU ELPA packages, something that has
> seen little progress in a long time, I think there is a case to be made
> for adding some of the more popular packages to core.

FWIW, my personal reason for using ace-window is due to helping with
RSI.  I find that the facility to navigate windows by staying on the
home row reduces the travel distance on the keyboard.

I recognise that this may not be as essential for other users, and offer
these remarks just for wider understanding.


From ace-window.el:
;; By default, ace-window uses numbers for window labels so the window
;; labeling is intuitively ordered.  But if you prefer to type keys on
;; your home row for quicker access, use this setting:
;;
;;    (setq aw-keys '(?a ?s ?d ?f ?g ?h ?j ?k ?l))


I've adapted it to Dvorak layout as:
(setq aw-keys '(?a ?s ?d ?f ?g ?h ?j ?k ?l))

Re: Suggestion: Ace-window in Emacs core for switching windows?

[Philip Kaludercic]

Jeremy Bryant <jb@jeremybryant.net> writes:

> Stefan Kangas <stefankangas@gmail.com> writes:
>
>> Philip Kaludercic <philipk@posteo.net> writes:
>>
>>> Given that we already have windmove, would the addition of ace-window
>>> really change that much?
>>
>> The same argument could be used for many packages that we have in core.
>> Who needs Gnus when you have rmail, and so on.  The packages do
>> basically the same thing, but in very different ways.

That is fair IMO, because the ways /are/ very different.  I just wanted
to clarify, if the method employed by ace-window is different enough to
be interesting as an alternative to windmove. 

>> I don't use ace-window myself, but note that it is the 28th most popular
>> package on MELPA with 1,773,799 downloads, only slightly behind
>> 'which-key' in popularity that we just added to core.
>>
>> See: https://melpa.org/#/?sort=downloads&asc=false

Popularity on MELPA correlates with age, especially since AFAIU they
count every upgrade as a download, so I am always hesitant to deduce too
much from data like these, especially when the popularity of a package
might be due to a missing feature in the core, that has since been added
or improved on.

>> Until we figure out how to bundle GNU ELPA packages, something that has
>> seen little progress in a long time, I think there is a case to be made
>> for adding some of the more popular packages to core.
>
> FWIW, my personal reason for using ace-window is due to helping with
> RSI.  I find that the facility to navigate windows by staying on the
> home row reduces the travel distance on the keyboard.
>
> I recognise that this may not be as essential for other users, and offer
> these remarks just for wider understanding.

My question is what your experience with windmove has been, and what
advantage ace-window has over windmove.  E.g. my main configuration for
windmove is

  (setopt windmove-default-keybindings '(nil . nil))

that binds window switching directly to the arrow keys.  I don't know of
anyone else who does this (and I know of a few colleges who got annoyed
at me for doing this), but I think it is neat.  What I am trying to say
is that there might be a windmove configuration that would serve you
just as well, but you just haven't found it yet.  In that case, we need
to document the features better, instead of going through the process of
adding more packages -- as we have seen, that can take a while.

> From ace-window.el:
> ;; By default, ace-window uses numbers for window labels so the window
> ;; labeling is intuitively ordered.  But if you prefer to type keys on
> ;; your home row for quicker access, use this setting:
> ;;
> ;;    (setq aw-keys '(?a ?s ?d ?f ?g ?h ?j ?k ?l))
>
>
> I've adapted it to Dvorak layout as:
> (setq aw-keys '(?a ?s ?d ?f ?g ?h ?j ?k ?l))

-- 
	Philip Kaludercic on peregrine

Re: Suggestion: Ace-window in Emacs core for switching windows?

[Stefan Kangas]

Philip Kaludercic <philipk@posteo.net> writes:

> Stefan Kangas <stefankangas@gmail.com> writes:
>
>> I don't use ace-window myself, but note that it is the 28th most popular
>> package on MELPA with 1,773,799 downloads, only slightly behind
>> 'which-key' in popularity that we just added to core.
>>
>> See: https://melpa.org/#/?sort=downloads&asc=false
>
> Popularity on MELPA correlates with age, especially since AFAIU they
> count every upgrade as a download, so I am always hesitant to deduce too
> much from data like these, especially when the popularity of a package
> might be due to a missing feature in the core, that has since been added
> or improved on.

All valid points, of course.  These statistics are far from perfect.
We also have users that download from GNU ELPA, directly from the
repository, etc.

At the same time, not all old packages reach such a high number of
downloads on MELPA -- some amount of popularity _is_ required.  As one
data point, one of my own packages is several years older than
ace-window, yet has a measly 3,000 downloads.  :-)

Re: Suggestion: Ace-window in Emacs core for switching windows?

[Jeremy Bryant]

Philip Kaludercic <philipk@posteo.net> writes:

> Jeremy Bryant <jb@jeremybryant.net> writes:
>
>> Stefan Kangas <stefankangas@gmail.com> writes:
>>
>>> Philip Kaludercic <philipk@posteo.net> writes:
>>>
>>>> Given that we already have windmove, would the addition of ace-window
>>>> really change that much?
>>>
>>> The same argument could be used for many packages that we have in core.
>>> Who needs Gnus when you have rmail, and so on.  The packages do
>>> basically the same thing, but in very different ways.
>
> That is fair IMO, because the ways /are/ very different.  I just wanted
> to clarify, if the method employed by ace-window is different enough to
> be interesting as an alternative to windmove. 


>> FWIW, my personal reason for using ace-window is due to helping with
>> RSI.  I find that the facility to navigate windows by staying on the
>> home row reduces the travel distance on the keyboard.
>>
>> I recognise that this may not be as essential for other users, and offer
>> these remarks just for wider understanding.
>
> My question is what your experience with windmove has been, and what
> advantage ace-window has over windmove.  E.g. my main configuration
> for

Happy to answer your question:
I used windmove before I started using ace-window, I concur that it is a
useful and intuitive package, and indeed more understandable.

I also like the fact that ace-window naturally scales up; in the case of
2 windows, M-x ace-window (bound to your key of choice), reduces to
simply the equivalent of M-x other-window.
For N>=3 windows, you get the interface explained before.

> windmove is
>
>   (setopt windmove-default-keybindings '(nil . nil))
>
> that binds window switching directly to the arrow keys.  I don't know of
> anyone else who does this (and I know of a few colleges who got annoyed
> at me for doing this), but I think it is neat.  What I am trying to say

I agree it's neat.
But in my personal use case, as explained, I wish to stay on the home row, so I
almost don't use the arrows for anything.

> is that there might be a windmove configuration that would serve you
> just as well, but you just haven't found it yet.  In that case, we need

Yes, it could well be true that there is a another interesting configuration.

> to document the features better, instead of going through the process of
> adding more packages -- as we have seen, that can take a while.

Yes, I accept the bar to adding new packages to Emacs is high

>
>> From ace-window.el:
>> ;; By default, ace-window uses numbers for window labels so the window
>> ;; labeling is intuitively ordered.  But if you prefer to type keys on
>> ;; your home row for quicker access, use this setting:
>> ;;
>> ;;    (setq aw-keys '(?a ?s ?d ?f ?g ?h ?j ?k ?l))
>>
>>
>> I've adapted it to Dvorak layout as:
>> (setq aw-keys '(?a ?s ?d ?f ?g ?h ?j ?k ?l))

Re: Suggestion: Ace-window in Emacs core for switching windows?

[Eli Zaretskii]

> From: Jeremy Bryant <jb@jeremybryant.net>
> Cc: ohwoeowho@gmail.com,  emacs-devel@gnu.org
> Date: Sat, 29 Jun 2024 23:10:33 +0100
> 
> Eli Zaretskii <eliz@gnu.org> writes:
> 
> >> Date: Tue, 25 Jun 2024 22:37:58 +0100
> >> From:  Jeremy Bryant via "Emacs development discussions." <emacs-devel@gnu.org>
> >> 
> >> 
> >> Ace-window is a quick way to switch windows, optionally using the home
> >> key row.
> >> 
> >> It has a high rating of 89% in ELPA
> >> https://elpa.gnu.org/packages/
> >> 
> >> I have found it useful and also stable - It doesn't appear to undergo
> >> major changes or need much ongoing maintenance.
> >> 
> >> Could it be a candidate for absorption into Emacs core?
> >
> > It could be, but why is it a problem to leave it on ELPA?
> 
> There are deployment scenarios where users can only use the Emacs
> provided in the distro, with no internet access for policy or security reasons.
> In particular, that means only built-in packages, without access to the internet.
> 
> As I personally use ace-window, I found it lacking in that scenario.

That argument can be applied to any ELPA package, can't it?

AFAIU, the lack of Internet access situation is no longer important
enough to warrant addition of packages to core.

(FTR, I need to work without Internet access quite a lot, so I'm
very much familiar with this situation and its problems.)

Re: Suggestion: Ace-window in Emacs core for switching windows?

[Jeremy Bryant]

Eli Zaretskii <eliz@gnu.org> writes:

>> > It could be, but why is it a problem to leave it on ELPA?
>> 
>> There are deployment scenarios where users can only use the Emacs
>> provided in the distro, with no internet access for policy or security reasons.
>> In particular, that means only built-in packages, without access to the internet.
>> 
>> As I personally use ace-window, I found it lacking in that scenario.
>
> That argument can be applied to any ELPA package, can't it?
>
> AFAIU, the lack of Internet access situation is no longer important
> enough to warrant addition of packages to core.

OK, I accept that

>
> (FTR, I need to work without Internet access quite a lot, so I'm
> very much familiar with this situation and its problems.)

Good to hear