path: root/2020/subtitles/emacsconf-2020--07-beyond-vim-and-emacs-a-scalable-ui-paradigm--sid-kasivajhula.vtt

WEBVTT

00:00:02.960 --> 00:00:04.644
"Far away in the heavenly abode

00:00:04.644 --> 00:00:06.560
of the great god Indra,

00:00:06.560 --> 00:00:07.688
there is a wonderful net

00:00:07.688 --> 00:00:10.160
which has been hung
by some cunning artificer

00:00:10.160 --> 00:00:12.080
in such a manner that it stretches out

00:00:12.080 --> 00:00:14.320
infinitely in all directions.

00:00:14.320 --> 00:00:16.938
In accordance with the extravagant
tastes of deities,

00:00:16.938 --> 00:00:18.240
the artificer has hung

00:00:18.240 --> 00:00:20.277
a single glittering jewel

00:00:20.277 --> 00:00:22.080
in each eye of the net,

00:00:22.080 --> 00:00:23.859
and since the net itself is infinite,

00:00:23.859 --> 00:00:26.480
the jewels are infinite in number.

00:00:26.480 --> 00:00:27.642
There hang the jewels,

00:00:27.642 --> 00:00:30.480
glittering like stars in the first
magnitude,

00:00:30.480 --> 00:00:32.681
a wonderful sight to behold.

00:00:32.681 --> 00:00:35.680
Were we to select one of these jewels
for inspection,

00:00:35.680 --> 00:00:38.216
we would discover that in
its polished surface

00:00:38.216 --> 00:00:39.520
there are reflected

00:00:39.520 --> 00:00:41.451
all the other jewels in the net,

00:00:41.451 --> 00:00:43.360
infinite in number.

00:00:43.360 --> 00:00:45.140
If we look still more closely,

00:00:45.140 --> 00:00:48.960
we would see that each of the jewels
reflected in this one jewel

00:00:48.960 --> 00:00:51.264
reflects all the others."

00:00:51.264 --> 00:00:54.000
This is the metaphor of Indra's Net,

00:00:54.000 --> 00:00:57.615
which is told in some schools of
philosophy.

00:00:57.615 --> 00:01:00.160
Let's keep this metaphor in mind,

00:01:00.160 --> 00:01:01.773
because it'll help us understand

00:01:01.773 --> 00:01:06.960
the Emacs extension that we're about to
discuss.

00:01:06.960 --> 00:01:12.810
In editing text, there's two main
paradigms:

00:01:12.810 --> 00:01:16.880
one is editing at the ground level,

00:01:16.880 --> 00:01:19.439
where the characters that we type

00:01:19.439 --> 00:01:22.159
actually appear on the screen,

00:01:22.159 --> 00:01:25.960
the changes we make actually occur.

00:01:28.479 --> 00:01:30.126
The other editing paradigm

00:01:30.126 --> 00:01:33.439
is where we escape to a higher level

00:01:33.439 --> 00:01:36.479
and now the characters that we type are
not...

00:01:36.479 --> 00:01:39.040
They don't actually appear on the screen

00:01:39.040 --> 00:01:42.748
because we're not at the ground level
with the text,

00:01:42.748 --> 00:01:44.799
we are at a higher level

00:01:44.799 --> 00:01:48.479
looking down at the text

00:01:48.479 --> 00:01:50.773
and regarding the text,

00:01:50.773 --> 00:01:56.159
referring to this world of text in terms
of a language.

00:01:56.159 --> 00:01:57.920
For instance, we could describe this

00:01:57.920 --> 00:02:03.404
world as having words and paragraphs and
sentences and lines and so on.

00:02:03.404 --> 00:02:05.985
We could reason about this text

00:02:05.985 --> 00:02:13.120
in terms of these textual entities and
this textual language.

00:02:13.120 --> 00:02:18.640
This is the second paradigm of text
editing.

00:02:18.640 --> 00:02:22.800
When we're in the second paradigm,

00:02:22.800 --> 00:02:25.304
there is a way to go down to ground
level.

00:02:25.304 --> 00:02:28.997
You hit Enter now--or we'll hit Enter to
go down to the ground level,

00:02:28.997 --> 00:02:30.480
and you can hit Escape

00:02:30.480 --> 00:02:33.200
to go back out to the referential level.

00:02:33.200 --> 00:02:35.200
Enter to go down to ground level

00:02:35.200 --> 00:02:40.160
and Escape to go up to the
referential level.

00:02:40.160 --> 00:02:47.565
Now, in Vim, the nouns in this
world of text

00:02:47.565 --> 00:02:52.319
all share the same referential plane
which we call normal mode.

00:02:52.319 --> 00:02:54.959
So in normal mode, all of the nouns

00:02:54.959 --> 00:02:57.360
of the world of text are available,

00:02:57.360 --> 00:03:00.959
whether it's words or sentences or
paragraphs,

00:03:00.959 --> 00:03:08.319
and they all share this same referential
plane.

00:03:08.319 --> 00:03:12.720
They compete for space on the keyboard.

00:03:12.720 --> 00:03:17.037
An alternative way to structure these
modes is

00:03:17.037 --> 00:03:21.840
instead of having a single mode where
all the nouns coexist,

00:03:21.840 --> 00:03:24.005
peacefully or otherwise,

00:03:24.005 --> 00:03:30.400
you instead have a dedicated mode for
every noun.

00:03:30.400 --> 00:03:32.540
In that case, what happens is

00:03:32.540 --> 00:03:35.440
because your modal spaces are
now much smaller,

00:03:35.440 --> 00:03:40.593
you're just talking about words or
paragraphs or lines or something,

00:03:40.593 --> 00:03:42.560
the keys that you use

00:03:42.560 --> 00:03:45.760
can be much more targeted.

00:03:45.760 --> 00:03:48.560
You can use the same keystrokes

00:03:48.560 --> 00:03:50.400
in all of your modes and they would have

00:03:50.400 --> 00:03:51.845
the same ideas behind them,

00:03:51.845 --> 00:03:53.280
but they would have different effects

00:03:53.280 --> 00:03:55.519
depending on which context you're using.

00:03:55.519 --> 00:03:59.120
It's the same keystrokes, different
contexts.

00:03:59.120 --> 00:04:04.244
The advantage of that is it's often
easier to change context

00:04:04.244 --> 00:04:07.888
than it is to learn new key bindings.

00:04:07.888 --> 00:04:11.289
So let's see an example of how
that works.

00:04:11.289 --> 00:04:14.039
We go into character mode, and if you
look at the mode line

00:04:14.039 --> 00:04:15.439
at the bottom of the screen there,

00:04:15.439 --> 00:04:18.720
you'll see that we're in character mode.

00:04:18.720 --> 00:04:21.955
Now, when we move up, down,
left, and right,

00:04:21.955 --> 00:04:23.919
we're moving by character.

00:04:23.919 --> 00:04:28.088
We can also transform the text,

00:04:28.088 --> 00:04:32.400
and the transformations occur in terms
of character.

00:04:32.400 --> 00:04:34.207
You can also go into word mode.

00:04:34.207 --> 00:04:40.000
In word mode, the transformations that
you do are on words.

00:04:40.000 --> 00:04:43.440
and you try... Your movement is also in
terms of words.

00:04:43.440 --> 00:04:46.560
So that's the level of granularity that
you have.

00:04:46.560 --> 00:04:49.191
You could also go to line mode.

00:04:49.191 --> 00:04:50.720
When you're in line mode,

00:04:50.720 --> 00:04:52.901
you go up and down by line,

00:04:52.901 --> 00:04:54.240
and you can move lines

00:04:54.240 --> 00:04:57.520
up and down left and right and so on.

00:04:59.120 --> 00:05:00.880
The transformations you do are in

00:05:00.880 --> 00:05:02.800
terms of lines.

00:05:02.800 --> 00:05:07.682
You could also go to window mode,

00:05:07.682 --> 00:05:11.695
where now the objects that you're
referring to are windows.

00:05:11.695 --> 00:05:15.578
You can move spatially
amongst the windows

00:05:15.578 --> 00:05:17.520
or do transformations on the windows

00:05:17.520 --> 00:05:20.850
using the same keystrokes.

00:05:25.360 --> 00:05:28.720
So let's go to...

00:05:28.720 --> 00:05:32.800
Right. One of the things,

00:05:32.800 --> 00:05:35.114
the principles at play here

00:05:35.114 --> 00:05:37.266
is something called the
Rumpelstiltskin principle,

00:05:37.266 --> 00:05:40.720
which is something that's known in
computer science.

00:05:40.720 --> 00:05:42.113
If you can name something,

00:05:42.113 --> 00:05:45.824
then you have power over it.

00:05:45.824 --> 00:05:48.560
This is kind of an adaptation of that
principle

00:05:48.560 --> 00:05:51.123
which says that if you can name
something

00:05:51.123 --> 00:05:52.572
and if you can talk about it,

00:05:52.572 --> 00:05:56.334
then it's a noun in your editing
language.

00:05:56.334 --> 00:05:58.960
If it's a noun, then it has...

00:05:58.960 --> 00:06:02.319
It's a mode. So if we can talk about it,
it's a noun.

00:06:02.319 --> 00:06:04.818
If it's a noun, then it's a mode.

00:06:04.818 --> 00:06:08.919
One of the things we've been talking a
lot about is modes.

00:06:08.919 --> 00:06:12.699
In fact, by this principle,

00:06:12.699 --> 00:06:17.280
modes also should be a mode.

00:06:17.280 --> 00:06:19.039
You should have a mode that can reason

00:06:19.039 --> 00:06:20.639
in terms of modes as objects,

00:06:20.639 --> 00:06:22.300
just like you have modes

00:06:22.300 --> 00:06:23.759
where you can reason in terms of

00:06:23.759 --> 00:06:26.560
words or lines as objects.

00:06:26.560 --> 00:06:30.479
So let's do that. Let's go to mode mode.

00:06:30.479 --> 00:06:34.000
When you go to mode mode, you see that

00:06:34.000 --> 00:06:35.915
the objects that are depicted here

00:06:35.915 --> 00:06:40.960
are the modes that are present
in the buffer,

00:06:40.960 --> 00:06:44.500
which we knew about because

00:06:44.500 --> 00:06:46.797
the style of editing that we had
in this buffer

00:06:46.797 --> 00:06:48.720
was the Vim style of editing

00:06:48.720 --> 00:06:51.143
where there's an insert mode at the
ground level

00:06:51.143 --> 00:06:53.039
and a normal mode that
you can escape to.

00:06:53.039 --> 00:06:57.280
You insert, enter the ground level.

00:06:57.280 --> 00:07:01.352
Enter to the insert mode and escape to
normal mode.

00:07:01.352 --> 00:07:04.647
When you look at the mode mode
representation,

00:07:04.647 --> 00:07:06.160
you see that in fact that

00:07:06.160 --> 00:07:10.479
is the structure that's depicted.

00:07:10.479 --> 00:07:14.080
But in different situations,
you might find

00:07:14.080 --> 00:07:16.080
that these modes are not the

00:07:16.080 --> 00:07:16.922
ones that you want.

00:07:16.922 --> 00:07:20.880
You want something more tailored for the
specific application.

00:07:20.880 --> 00:07:25.065
For instance, if you're editing
Lisp code

00:07:25.065 --> 00:07:27.360
(or code in general, but

00:07:27.360 --> 00:07:30.880
Lisp code is a particular example),

00:07:30.880 --> 00:07:32.640
you might want to take advantage

00:07:32.640 --> 00:07:34.852
of the structure of the code.

00:07:34.852 --> 00:07:37.599
For Lisp code in particular,

00:07:37.599 --> 00:07:40.960
we have a mode called symex-mode

00:07:40.960 --> 00:07:45.414
which is able to reason about your code

00:07:45.414 --> 00:07:47.919
in terms of its tree structure.

00:07:47.919 --> 00:07:52.397
So you can use the same keystrokes: hjkl
goes left, right, up, and down,

00:07:52.397 --> 00:07:58.080
but you also have other keystrokes that
are more specialized to the application.

00:07:58.080 --> 00:08:01.520
You can run the code.

00:08:01.520 --> 00:08:06.960
We'll see that happen here in a minute.

00:08:06.960 --> 00:08:12.240
You can make changes to it
really quickly

00:08:12.240 --> 00:08:18.000
and see the effects of those changes.

00:08:18.000 --> 00:08:19.440
You're doing this all in a mode

00:08:19.440 --> 00:08:22.625
that's convenient for this particular
application,

00:08:22.625 --> 00:08:25.039
which is editing Lisp code,

00:08:25.039 --> 00:08:28.960
and that is, in this case, symex-mode.

00:08:28.960 --> 00:08:31.631
Typically, when you're editing code
like this,

00:08:31.631 --> 00:08:33.435
you'd want to be in insert mode

00:08:33.435 --> 00:08:36.640
actually typing out the code,

00:08:36.640 --> 00:08:40.959
and then you'd want to escape to symex
mode rather than normal mode,

00:08:40.959 --> 00:08:42.021
and then you could escape again

00:08:42.021 --> 00:08:44.080
and you'd end up in normal mode.

00:08:44.080 --> 00:08:48.000
So this, if we go to mode mode, we see
is depicted

00:08:48.000 --> 00:08:51.040
as this tower where insert is at the

00:08:51.040 --> 00:08:52.604
bottom and normal is at the top,

00:08:52.604 --> 00:08:55.305
but symex-mode is in between the two.

00:08:55.305 --> 00:08:57.551
You could also change that if you like.

00:08:57.551 --> 00:08:59.566
If you don't want symex-mode
to be there,

00:08:59.566 --> 00:09:02.187
you could just move it to the top.

00:09:02.187 --> 00:09:04.392
Now you find symex is at the top

00:09:04.392 --> 00:09:06.160
and you enter down to normal.

00:09:06.160 --> 00:09:08.848
You can see it on the status bar at the
bottom there.

00:09:08.848 --> 00:09:13.839
Enter to insert, escape to normal,
escape to symex.

00:09:13.839 --> 00:09:16.344
In fact, you can even add more modes

00:09:16.344 --> 00:09:19.380
if you don't like the existing ones.

00:09:21.519 --> 00:09:23.839
Now we have an additional mode here.

00:09:23.839 --> 00:09:25.855
We have window mode. It goes
down to symex,

00:09:25.855 --> 00:09:27.519
it goes down to normal.

00:09:27.519 --> 00:09:29.919
Enter the insert, escape to normal,

00:09:29.919 --> 00:09:33.600
escape to symex, escape to window.

00:09:33.600 --> 00:09:41.232
So we've talked... Okay, so another thing
actually to note here

00:09:41.232 --> 00:09:45.360
is that in editing modes,

00:09:45.360 --> 00:09:46.486
if you look at the mode line

00:09:46.486 --> 00:09:48.399
at the bottom of the screen,

00:09:48.399 --> 00:09:50.257
you'll see that we are currently,

00:09:50.257 --> 00:09:51.519
in this buffer,

00:09:51.519 --> 00:09:54.560
we are currently in line mode.

00:09:54.560 --> 00:09:57.296
I'm going to hit Enter now

00:09:57.296 --> 00:09:59.119
and you'll see that when I hit Enter,

00:09:59.119 --> 00:10:00.627
nothing is happening.

00:10:00.627 --> 00:10:02.160
It's still in line mode.

00:10:02.160 --> 00:10:05.120
If you hit Escape, it's still in line mode.

00:10:05.120 --> 00:10:07.200
You can find out the reason for that

00:10:07.200 --> 00:10:10.640
by taking another meta jump out of this.

00:10:10.640 --> 00:10:12.800
You'll see that, in fact, the reason

00:10:12.800 --> 00:10:15.279
is that we're currently in line mode,

00:10:15.279 --> 00:10:17.360
and line mode is the only one available

00:10:17.360 --> 00:10:19.519
in this tower

00:10:19.519 --> 00:10:21.556
for editing the modes that are

00:10:21.556 --> 00:10:24.880
in operation in your ground level.

00:10:24.880 --> 00:10:26.898
In fact, line mode is all you need
here,

00:10:26.898 --> 00:10:32.796
because this is just the nature of how
these modes are laid out is in rows.

00:10:32.796 --> 00:10:36.399
So line mode is the most appropriate
thing here.

00:10:36.399 --> 00:10:39.740
But you could change it to something
else if you like.

00:10:40.959 --> 00:10:43.659
Now we've seen two towers.

00:10:43.659 --> 00:10:53.680
We've seen the Vim tower and we've seen
also the symex tower, the Lisp tower.

00:10:53.680 --> 00:10:58.880
It turns out that, because we've been
talking about towers now,

00:10:58.880 --> 00:11:06.399
by the Rumpelstiltskin principle, towers
also can be talked about,

00:11:06.399 --> 00:11:09.127
and therefore they also are a mode.

00:11:09.127 --> 00:11:11.200
So how do we go to tower mode?

00:11:11.200 --> 00:11:14.640
The way we go to tower mode is

00:11:14.640 --> 00:11:19.200
we go in a slightly different direction,

00:11:19.200 --> 00:11:23.360
and we find that we are now in tower
mode.

00:11:23.360 --> 00:11:29.279
We see that there are many towers
available. We're now...

00:11:29.279 --> 00:11:33.440
We're seeing several possible towers

00:11:33.440 --> 00:11:40.344
that we have written to be available and
for use in different buffers.

00:11:40.344 --> 00:11:42.110
You can edit them on the fly.

00:11:42.110 --> 00:11:46.630
For instance, let's enter this tower.

00:11:48.000 --> 00:11:50.180
Now you see that in the bottom of
the...

00:11:50.180 --> 00:11:51.519
In the mode line, you see that we're

00:11:51.519 --> 00:11:53.944
going across all of these
different modes

00:11:53.944 --> 00:11:56.480
that were in the tower.

00:11:56.480 --> 00:11:59.724
You could escape and you could even move
things around.

00:11:59.724 --> 00:12:00.880
You could put window mode

00:12:00.880 --> 00:12:02.573
all the way at the bottom,

00:12:02.573 --> 00:12:04.079
right above insert mode.

00:12:04.079 --> 00:12:06.479
Let's see that happen. There it is,

00:12:06.479 --> 00:12:10.444
window is right above insert, and
so on.

00:12:10.444 --> 00:12:14.240
The tower always reflects your current
position,

00:12:14.240 --> 00:12:17.600
so if you're in buffer mode here and you
go down to line mode,

00:12:17.600 --> 00:12:22.480
when you go back to mode mode, you see
that we are in line mode.

00:12:22.480 --> 00:12:25.620
But in practice, you wouldn't have a
tower this elaborate

00:12:25.620 --> 00:12:29.440
because you'd rather have several
smaller towers you enter,

00:12:29.440 --> 00:12:33.360
that you alternate between.

00:12:33.360 --> 00:12:39.839
Okay. So one other thing of interest
here is that

00:12:39.839 --> 00:12:42.240
when you're in tower mode,

00:12:42.240 --> 00:12:44.740
if you look at the status line at the
bottom there,

00:12:44.740 --> 00:12:49.839
we are currently in buffer mode while we
are in tower mode.

00:12:49.839 --> 00:12:53.151
Tower mode actually isn't a mode really.
Neither is mode mode.

00:12:53.151 --> 00:12:58.000
They're really referential planes or
meta planes.

00:12:58.000 --> 00:13:01.679
In any case, you can see that we're in
buffer mode.

00:13:01.679 --> 00:13:03.840
We can take a meta jump out of this

00:13:03.840 --> 00:13:08.000
to confirm that buffer mode is the only
mode available

00:13:08.000 --> 00:13:09.664
when we're editing towers

00:13:09.664 --> 00:13:11.915
because that's the one we need,

00:13:11.915 --> 00:13:23.200
given that our towers are represented in
individual buffers.

00:13:23.200 --> 00:13:26.320
Right. So let's see where we're at.

00:13:26.320 --> 00:13:27.785
Rumpelstiltskin principle...

00:13:27.785 --> 00:13:30.160
We talked about mode mode.

00:13:30.160 --> 00:13:32.240
We talked about the strange loop

00:13:32.240 --> 00:13:37.820
application of ground level modes in
meta levels.

00:13:39.600 --> 00:13:41.992
We saw the different towers,

00:13:41.992 --> 00:13:50.720
and in fact, we're currently
in Vim tower,

00:13:50.720 --> 00:13:52.860
where you can go to Emacs tower.

00:13:52.860 --> 00:13:54.720
Now, with a single keystroke, you can

00:13:54.720 --> 00:13:59.695
alternate between Emacs and Vim,

00:13:59.695 --> 00:14:01.638
which are represented--

00:14:01.638 --> 00:14:05.519
which are modeled as towers.

00:14:13.360 --> 00:14:14.760
So there's... One thing

00:14:14.760 --> 00:14:18.160
that we've sort of alluded to is that
there are two directions

00:14:18.160 --> 00:14:19.494
that you can travel in

00:14:19.494 --> 00:14:22.399
when you're going through this
framework.

00:14:22.399 --> 00:14:33.760
One direction is--and we'll visualize it
like so...

00:14:33.760 --> 00:14:35.120
There's two directions you can travel,

00:14:35.120 --> 00:14:37.040
and you can either go sideways or you

00:14:37.040 --> 00:14:38.399
can go up and down.

00:14:38.399 --> 00:14:41.680
If you go sideways, you're changing your
perspective.

00:14:41.680 --> 00:14:45.440
So normal mode, word mode, line mode,

00:14:45.440 --> 00:14:46.544
window mode, and so on

00:14:46.544 --> 00:14:51.680
are all different perspectives on your
ground editing experience.

00:14:51.680 --> 00:14:53.265
The other direction you can travel in

00:14:53.265 --> 00:14:56.811
is up or down, which takes you through
meta levels.

00:14:56.811 --> 00:14:59.600
So you go from the ground level
editing experience,

00:14:59.600 --> 00:15:07.040
up to mode mode, and then up to the
tower plane, and so on, and so on.

00:15:07.040 --> 00:15:12.568
So this all sounds very complex,

00:15:12.568 --> 00:15:18.160
but the truth is it's not really that
complicated,

00:15:18.160 --> 00:15:20.699
even though it feels that way.

00:15:20.699 --> 00:15:22.959
The reason it isn't that complicated

00:15:22.959 --> 00:15:26.480
is because no matter how many levels

00:15:26.480 --> 00:15:30.160
up or down you go and no matter where
you are,

00:15:30.160 --> 00:15:32.399
whether you're in at the ground level

00:15:32.399 --> 00:15:34.079
editing the actual text

00:15:34.079 --> 00:15:35.802
or whether you're at a meta level,

00:15:35.802 --> 00:15:39.279
some unknown meta level and you don't
know where you are,

00:15:39.279 --> 00:15:41.133
no matter where you are,

00:15:41.133 --> 00:15:44.399
the way in which you interact with it

00:15:44.399 --> 00:15:47.519
is the same at every level.

00:15:47.519 --> 00:15:54.751
That is the great power of this
approach:

00:15:54.751 --> 00:16:00.880
that all of the different levels
are the same.

00:16:00.880 --> 00:16:03.839
In fact, the complexity of the whole

00:16:03.839 --> 00:16:05.545
is exactly identical to

00:16:05.545 --> 00:16:07.657
the complexity of each part,

00:16:07.657 --> 00:16:10.000
so if you know how to edit words

00:16:10.000 --> 00:16:12.048
in the ground level buffer

00:16:12.048 --> 00:16:15.378
and you know how to move lines around
using line mode,

00:16:15.378 --> 00:16:22.800
then you know how to edit any aspect of
your editing experience at any level.

00:16:30.079 --> 00:16:31.780
So this is a pre-release demo.

00:16:31.780 --> 00:16:40.079
This doesn't exist on MELPA yet, but you
can follow updates at this repo on
github.

00:16:40.079 --> 00:16:43.850
If you can also be a beta tester

00:16:43.850 --> 00:16:46.775
or something like that, if you like,
that would be very helpful.

00:16:46.775 --> 00:16:50.560
You can learn more about this at

00:16:50.560 --> 00:16:53.920
drym.org, which is where I house

00:16:53.920 --> 00:16:55.726
the research that I work on.

00:16:55.726 --> 00:17:00.154
In particular, the research on epistemic
levels

00:17:00.154 --> 00:17:03.600
is what inspired this particular Emacs
extension.

00:17:03.600 --> 00:17:05.600
You can also learn about

00:17:05.600 --> 00:17:10.880
dialectical inheritance attribution,
which is the basis of

00:17:10.880 --> 00:17:14.559
a new economic system that could be fair

00:17:14.559 --> 00:17:19.439
and could lead to a prosperous and happy
world.

00:17:19.439 --> 00:17:26.319
You can follow me on Twitter at
@countvajhula.

00:17:26.319 --> 00:17:31.919
That's it! Thank you.