Re: [O] Emulating list functionality from traditional GUI editors

[Jorge A. Alfaro-Murillo]

Calvin Young writes:

> ;; Starting with this setup: 
>
> - one - two + a[CURSOR_HERE] 
>
> ;; Hitting <enter> should produce: 
>
> - one - two + a + [CURSOR_HERE] 
>
> ;; Hitting <enter> again would then produce: 
>
> - one - two + a - [CURSOR_HERE] 
>
> ;; And hitting <enter> one last time would produce: 
>
> - one - two + a 
>
> [CURSOR_HERE] 

I think there is a confusion here, my understanding is that org 
separates sublists by indentation so if you have:

- a
+ b[CURSOR]

and hit M-<enter> it should correct to:

- a
- b
- [CURSOR]

It is different if you have:

- a
 + b[CURSOR]

or

- a
 - b[CURSOR]

or 

1. a
  - b[CURSOR]
2. c

etc

> I know we can already achieve this with some combination of 
> M-<enter>, <enter>, and M-S-<enter>, but this behavior has 2 
> distinct advantages: 
>
> 1. The user only needs to remember one key to cycle between all 
> of these actions, rather than 3 key combinations.

But the problem is that you lose the functionality of <enter> to 
exit the list. I want to have <enter> to finish a line and <enter> 
<enter> to finish a paragraph like I am used everywhere else.

> 2. This behavior is more consistent with the bulleting behavior 
> in other editors

We shouldn't aim to imitate other much inferior editors ;-)

> , which could make it feel more intuitive for new org-mode 
> users.  [...]
> Yes, but for the reasons mentioned above, it'd be nice if we 
> could use the <enter> to outdent a new list entry as well. 

I disagree, <tab> and S-<tab> for indenting is much more friendly 
than <enter>, most modes in emacs behave like that. A new user 
just has to understand that sublists are separated by indentation, 
and learn that M-<enter> is for lists, and M-S-<enter> is for 
check boxes.

> Jorge A. Alfaro-Murillo says:
> > I guess you could remap <backspace> to a function that checks 
> > if you are at the beginning of the list and when that is true 
> > it does what you want, otherwise it just calls 
> > `delete-backward-char'. But generally I would do C-a C-k 
> > <backspace>, just two more keystrokes.
> Makes sense. This is an easy function to write — just wanted to 
> make sure there wasn't something that already does this 
> out-of-the-box.

It should save around half a second per use, so if you use it five 
times a day you have about 1 hour to write it... minus the time it 
takes you to read this: http://xkcd.com/1205/ =)

> In general, I *believe* a lot of folks use lists and checkboxs 
> in similar ways. I certainly do, and I frequently accidentally 
> hit M-<enter> while editing a checkbox when I really intend to 
> insert a new checkbox entry. As a result, it seems desirable to 
> create an interface that treats them more similarly (e.g., using 
> a single <enter> keypress to auto-insert a new entry).
>
> If this doesn't exist yet, I'd be happy to roll it myself. But 
> it'd be nice to avoid re-inventing the wheel here if possible :) 

Maybe, but you would lose the ability to have mixed check boxes 
and items lists. But you are right, it might be a nice 
configuration to allow M-<enter> to give you another item with 
check box if you are already in one (and then M-S-<enter> gives 
you a plain list item). But it might be even more confusing for a 
new user as to why the behavior is not consistent with M-<enter>, 
so probably it shouldn't be default.

Best,

--
Jorge.