path: root/2021/captions/emacsconf-2021-bidi--perso-arabic-input-methods-and-making-more-emacs-apps-bidi-aware--mohsen-banan--main.vtt

WEBVTT

00:02.960 --> 00:00:10.240
Greetings. Salaam. This is Mohsen Banan. 

00:00:10.240 --> 00:00:14.080
I am an Iranian software and internet engineer.

00:14.080 --> 00:17.199
I converted to Emacs in 1986.

00:17.199 --> 00:20.480
It was Emacs version 17 then.

00:20.480 --> 00:00:23.920
By around 1988, when Emacs version 18

00:00:23.920 --> 00:00:25.680
was well in place, 

00:00:25.680 --> 00:00:28.960
I started living inside of Emacs. 

00:00:28.960 --> 00:00:31.119
My primary digital environment 

00:00:31.119 --> 00:00:33.680
has been Emacs ever since.

00:33.680 --> 00:36.559
It has been a good life.

00:36.559 --> 00:40.000
I'm a native Farsi speaker and writer.

00:40.000 --> 00:00:41.520
I'm not a linguist, 

00:00:41.520 --> 00:00:43.800
and I do not specialize in

00:43.800 --> 00:46.719
multilingualization, internationalization,

00:46.719 --> 00:48.879
and localization.

00:48.879 --> 00:52.239
My favorite programming language is Lisp,

00:52.239 --> 00:00:58.319
and I am a bit of an Emacs developer.

00:58.320 --> 01:00.719
This presentation is about use of

01:00.719 --> 01:04.000
Perso-Arabic Scripts with Emacs.

01:04.000 --> 01:06.720
It's an overview presentation.

01:06.720 --> 00:01:08.320
I won't be digging deep 

00:01:08.320 --> 00:01:11.360
in many of the mentioned topics.

01:11.360 --> 00:01:13.119
My goal is to make you aware 

00:01:13.119 --> 00:01:16.159
of what can be done with Emacs today,

01:16.159 --> 01:19.280
and the potentials that Emacs presents

01:19.280 --> 01:22.640
for Perso-Arabic writers.

01:22.640 --> 01:25.759
The main topics that i'll cover are:

01:25.759 --> 00:01:30.320
- a brief introduction to Perso-Arabic scripts

01:30.320 --> 01:35.119
- two existing Emacs Persian input methods

01:35.119 --> 00:01:36.960
- the challenges involved with

00:01:36.960 --> 00:01:42.960
making Emacs applications bidirectional-aware

01:42.960 --> 00:01:45.280
- the ultimate goal of creating 

00:01:45.280 --> 00:01:47.759
a complete digital environment 

00:01:47.759 --> 00:01:51.200
for Perso-Arabic writers

01:51.200 --> 00:01:57.679
I'll also be including various pointers.

01:57.680 --> 00:02:00.320
So first, let's make sure 

00:02:00.320 --> 00:02:02.079
that what I'm presenting 

00:02:02.079 --> 00:02:04.560
is of interest to you.

02:04.560 --> 00:02:07.040
if you are a Perso-Arabic writer 

00:02:07.040 --> 00:02:08.879
and if you use Emacs, 

00:02:08.879 --> 00:02:12.720
you're definitely my intended audience.

02:12.720 --> 00:02:15.120
If you're an Emacs developer 

00:02:15.120 --> 00:02:19.120
who wishes to make her Emacs apps

02:19.120 --> 00:02:22.000
multilingual and bidi-aware, 

00:02:22.000 --> 00:02:27.359
you're also my intended audience.

02:27.360 --> 00:02:30.480
For the purposes of this presentation, 

00:02:30.480 --> 00:02:34.160
in this slide, I'm categorizing scripts

02:34.160 --> 02:39.040
based on directionality and shaping.

02:39.040 --> 00:02:41.280
Latin letters are not shaped. 

00:02:41.280 --> 00:02:43.120
Generally speaking, 

00:02:43.120 --> 00:02:44.800
the shape of a Latin letter 

00:02:44.800 --> 00:02:49.599
is independent of its position in a word.

02:49.599 --> 00:02:53.280
Perso-Arabic letters are subject to shaping.

02:53.280 --> 02:56.319
For example the letter mim--

02:56.319 --> 02:58.800
sounding similar to M--

02:58.800 --> 03:01.840
takes three shapes depending on whether

03:01.840 --> 03:04.720
it is in the beginning of a word,

03:04.720 --> 00:03:06.480
in the middle of a word, 

00:03:06.480 --> 00:03:09.200
or at the end of a word.

03:09.200 --> 00:03:12.159
I'll be showing more of how shaping works

00:03:12.159 --> 00:03:16.400
in an Emacs session screencast later.

03:16.400 --> 00:03:18.159
Shaping has ramifications 

00:03:18.159 --> 00:03:21.040
for Emacs application developers.

03:21.040 --> 00:03:23.760
For example, if you are combining

00:03:23.760 --> 00:03:26.720
initial letters to create a label,

03:26.720 --> 03:29.680
those letters can be shaped together--

03:29.680 --> 03:31.760
which is not what you want.

03:31.760 --> 03:34.000
In such cases, you would need to

03:34.000 --> 03:37.840
explicitly keep them separate.

03:37.840 --> 00:03:40.799
Latin-based scripts are always 

00:03:40.799 --> 00:03:42.560
left-to-right.

03:42.560 --> 00:03:46.560
Perso-Arabic scripts are right-to-left

03:46.560 --> 00:03:51.040
with letters, but numbers are left-to-right.

03:51.040 --> 00:03:53.040
So, Perso-Arabic scripts 

00:03:53.040 --> 00:03:56.319
are bi-directional (BIDI).

03:56.319 --> 00:03:58.879
Hebrew is also bi-directional, 

00:03:58.879 --> 00:04:02.159
but Hebrew is not shaped.

04:02.159 --> 00:04:05.680
More recently, it has become very common 

00:04:05.680 --> 00:04:10.400
to mix Perso-Arabic and Latin text.

04:10.400 --> 00:04:12.560
This can become very confusing 

00:04:12.560 --> 00:04:14.959
if paragraph directionality 

00:04:14.959 --> 00:04:17.759
is not properly observed.

04:17.759 --> 00:04:20.000
I'll be providing some examples 

00:04:20.000 --> 00:04:23.120
as screencasts.

04:23.120 --> 00:04:24.880
The Emacs display engine 

00:04:24.880 --> 00:04:27.280
now fully and well supports 

00:04:27.280 --> 00:04:32.079
both shaping and BIDI.

04:32.080 --> 00:04:37.120
Since 2012, starting with Emacs version 24,

04:37.120 --> 00:04:38.639
we can say that Emacs 

00:04:38.639 --> 00:04:43.600
is a truly multilingual-capable environment.

04:43.600 --> 00:04:45.040
Like everything else,

00:04:45.040 --> 00:04:46.880
multilingual support for Emacs 

00:04:46.880 --> 00:04:49.280
was added gradually.

04:49.280 --> 04:52.560
Unicode support was added early on.

04:52.560 --> 04:55.520
The framework for input methods evolved

04:55.520 --> 00:05:00.320
in the 1990s. But it was not till version 24 

00:05:00.320 --> 00:05:03.360
in 2012 that the display engine 

00:05:03.360 --> 00:05:06.240
could fully support BIDI.

05:06.240 --> 00:05:08.800
Hats off to Eli Zaretskii 

00:05:08.800 --> 00:05:12.960
for his work on Emacs BIDI.

05:12.960 --> 00:05:15.600
Once full BIDI support was in place

00:05:15.600 --> 00:05:19.199
in 2012, I went ahead and added

00:05:19.199 --> 00:05:24.000
two Persian input methods to Emacs 24.

05:24.000 --> 00:05:26.400
So now Emacs fully supports 

00:05:26.400 --> 00:05:30.638
Perso-Arabic scripts.

05:30.639 --> 00:05:34.080
By Perso-Arabic script, we're referring to

00:05:34.080 --> 00:05:36.320
the Arabic writing system

05:36.320 --> 00:05:38.160
with various extensions 

00:05:38.160 --> 00:05:42.240
used by a large number of languages.

05:42.240 --> 00:05:44.479
Perso-Arabic is the second most 

00:05:44.479 --> 00:05:47.520
widely used writing system in the world

05:47.520 --> 05:50.479
by the number of countries.

05:50.479 --> 05:54.000
It is the third by the number of users

05:54.000 --> 05:59.039
after the Latin and Chinese scripts.

05:59.039 --> 06:02.160
So, by well supporting Perso-Arabic,

06:02.160 --> 00:06:04.720
Emacs's potential user base 

00:06:04.720 --> 00:06:09.918
can be greatly enhanced.

06:09.919 --> 00:06:12.960
Before focusing on the Persian input methods,

00:06:12.960 --> 00:06:14.960
let me quickly summarize 

00:06:14.960 --> 00:06:18.319
Emacs's input methods model.

06:18.319 --> 00:06:21.360
Input methods allow you to enter characters 

00:06:21.360 --> 00:06:25.120
that are not supported by your keyboard.

06:25.120 --> 00:06:29.280
With Quail maps, we can map ASCII key strings 

00:06:29.280 --> 00:06:31.919
to multilingual characters.

06:31.919 --> 00:06:34.080
So we can input any text 

00:06:34.080 --> 00:06:36.960
from an ASCII keyboard.

06:36.960 --> 00:06:38.960
You select an input method 

00:06:38.960 --> 00:06:44.400
with C-x RET C-\ .

06:44.400 --> 00:06:49.918
We'll try that in a screencast shortly.

06:49.919 --> 00:06:53.919
Since version 24, Emacs comes loaded with

00:06:53.919 --> 00:06:56.880
two Persian input methods:

06:56.880 --> 00:07:01.520
farsi-isiri-9147 is the standard

00:07:01.520 --> 00:07:04.960
traditional Iranian keyboard.

07:04.960 --> 00:07:09.199
farsi-transliterate-banan is an intuitive

00:07:09.199 --> 00:07:12.319
transliteration keyboard for Farsi

00:07:12.319 --> 00:07:16.960
which requires near-zero training for use.

07:16.960 --> 00:07:18.960
I'll be mostly focused on 

00:07:18.960 --> 00:07:21.520
farsi-transliterate-banan 

00:07:21.520 --> 00:07:23.919
in this presentation.

07:23.919 --> 00:07:27.839
So let's try this out.

07:27.840 --> 07:30.319
In this gif-cast, we're going to select

07:30.319 --> 00:07:32.240
a Persian input method 

00:07:32.240 --> 00:07:35.280
and write a few simple sentences.

07:35.280 --> 07:38.160
With no training and no documentation,

07:38.160 --> 00:07:41.360
any Farsi writer familiar with Emacs 

00:07:41.360 --> 00:07:45.440
can write these, as farsi-transliterate-banan 

00:07:45.440 --> 00:07:48.560
input method is intuitive.

07:48.560 --> 00:07:50.400
I'll be using keycast 

00:07:50.400 --> 00:07:54.160
to show you keys as they are used.

07:54.160 --> 07:56.560
Let me first describe as to what we have

07:56.560 --> 07:58.720
on the screen.

07:58.720 --> 08:01.520
There are three windows in one frame.

08:01.520 --> 00:08:04.639
Keycast will show commands and keys 

00:08:04.639 --> 00:08:06.319
on the mode line.

08:06.319 --> 00:08:08.639
The leftmost window is showing 

00:08:08.639 --> 00:08:10.879
logs of keycast.

08:10.879 --> 08:14.639
Transformed individual unshaped letters

08:14.639 --> 08:17.360
will appear here.

08:17.360 --> 00:08:21.120
The middle window is running a tail -f 

00:08:21.120 --> 00:08:27.440
on the dribble file piped to fold -w1.

08:27.440 --> 00:08:30.720
This lets you see the raw ASCII characters

00:08:30.720 --> 00:08:34.479
as I type them.

08:34.479 --> 00:08:37.839
The right window is the empty buffer 

00:08:37.839 --> 00:08:41.680
on the ex.fa file.

08:41.680 --> 00:08:44.560
Anything that I described here 

00:08:44.560 --> 00:08:49.040
can be done with virgin Emacs distribution

08:49.040 --> 08:51.360
with nothing added,

08:51.360 --> 08:53.760
but I'm using Blee

08:53.760 --> 08:57.120
(By* Libre-Halaal Emacs Environment)

08:57.120 --> 00:09:02.720
to show things. You don't need to have blee

00:09:02.720 --> 00:09:07.600
for writing the equivalent of the text 

00:09:07.600 --> 00:09:10.800
in this gif-cast.

09:10.800 --> 00:09:13.440
First, I'm going to select 

00:09:13.440 --> 00:09:16.800
the farsi-transliterate-banan.

09:16.800 --> 00:09:22.080
I'm entering C-x RET C-\ .

09:22.080 --> 00:09:23.839
Notice the mode -ine 

00:09:23.839 --> 00:09:27.040
and the prompt at mini-buffer.

09:27.040 --> 09:30.399
With completion, I'm going to select

09:30.399 --> 09:34.000
farsi-transliterate-banan.

09:34.000 --> 00:09:38.560
Notice that farsi-isiri-9147

09:38.560 --> 09:43.360
was also provided as a choice.

09:43.360 --> 00:09:47.440
Also notice that the letter 'b' appears

00:09:47.440 --> 00:09:51.760
in the left of the mode line of ex.fa.

09:51.760 --> 00:09:54.000
This indicates which input method 

00:09:54.000 --> 00:09:56.320
has been selected.

09:56.320 --> 00:09:58.560
Also notice that cursor is on 

00:09:58.560 --> 00:10:04.399
the top left corner of ex.fa.

10:04.399 --> 10:08.800
Next, I'm going to enter the 's' character.

10:08.800 --> 10:12.000
Notice the cursor moved to the right,

10:12.000 --> 00:10:15.120
and unshaped ’seen’ appeared 

00:10:15.120 --> 00:10:20.079
in the ex.fa buffer, on the mode-line,

10:20.079 --> 10:23.760
and in the keycast log buffer.

10:23.760 --> 00:10:28.000
Next, I'm going to enter 'l'.

10:28.000 --> 00:10:32.800
Notice ل in the mode-line,

10:32.800 --> 00:10:39.920
and notice how س was subjected to shaping.

10:39.920 --> 10:42.640
Next I'm going to write the letter

10:42.640 --> 10:44.640
to write the following:

10:44.640 --> 10:50.560
 سلام –حال شما چه طوره؟ –با ایمکس همه کار میشه کرد

10:50.560 --> 10:52.480
"Hello, how are you?"

10:52.480 --> 00:10:56.720
You can do everything with Emacs.

10:56.720 --> 11:00.000
Generally, same-sounding Latin characters

11:00.000 --> 00:11:03.920
are used. As usual, vowels are ignored

00:11:03.920 --> 00:11:06.720
unless called for.

11:06.720 --> 11:10.800
Notice that in order to get ح,

11:10.800 --> 11:14.560
i repeated 'h' twice.

11:14.560 --> 11:15.440
ش

11:15.440 --> 11:18.560
is the obvious 'sh'.

11:18.560 --> 11:19.519
چ

11:19.519 --> 11:22.480
is the obvious 'ch'.

11:22.480 --> 00:11:28.640
t, that's the ط  is upper case T.

11:28.640 --> 00:11:29.439
 ت 

00:11:29.439 --> 00:11:31.839
is lowercase t.

11:31.839 --> 11:34.880
That's it. We managed to write in Farsi

11:34.880 --> 11:38.640
with a QWERTY keyboard, intuitively.

11:38.640 --> 11:40.959
Next, we are going to switch back to

11:40.959 --> 00:11:45.680
globish and write "back to globish".

11:45.680 --> 00:11:47.920
Notice that the globish sentence

00:11:47.920 --> 00:11:50.959
started from the left side.

11:50.959 --> 00:11:53.279
This is due to proper detection 

00:11:53.279 --> 00:12:00.159
of paragraph directionality by Emacs.

12:00.160 --> 00:12:05.600
For the most part, Emacs is self-documenting.

12:05.600 --> 12:07.920
Here we are pointing you to some

12:07.920 --> 12:12.880
relevant self-contained Emacs resources.

12:12.880 --> 00:12:15.680
The BIDI documentation applies to

00:12:15.680 --> 00:12:18.000
all BIDI scripts,

12:18.000 --> 12:22.079
not just Perso-Arabic scripts.

12:22.079 --> 00:12:23.279
Referring to the code 

00:12:23.279 --> 00:12:27.999
can also be useful for some.

12:28.000 --> 12:30.079
Here are some pointers.

12:30.079 --> 00:12:32.399
The Quail translation code 

00:12:32.399 --> 00:12:35.519
for Persian input methods: 

00:12:35.519 --> 00:12:39.120
the persian.el file has full details 

00:12:39.120 --> 00:12:43.999
of the mapping and some documentation.

12:44.000 --> 00:12:46.639
Next, we'll show the keyboard layouts 

00:12:46.639 --> 00:12:51.519
as a gif-cast. You can get

00:12:51.519 --> 00:12:55.519
relevant documentation for any input method

12:55.519 --> 00:13:00.320
with the describe-input-method command.

13:00.320 --> 00:13:05.920
So, let's try that for farsi-transliterate-banan.

00:13:05.920 --> 00:13:09.360
We are back in the ex.fa buffer 

00:13:09.360 --> 00:13:11.279
as one window.

13:11.279 --> 13:13.839
We don't need the keycast logging

13:13.839 --> 13:17.040
and the dribble windows any more.

13:17.040 --> 13:20.320
With the C-\, I reactivate

13:20.320 --> 13:23.360
the farsi input method.

13:23.360 --> 00:13:26.399
Notice that keycast is still active 

00:13:26.399 --> 00:13:29.360
on the mode line.

13:29.360 --> 00:13:34.240
Next, with the C-h C-\ , 

00:13:34.240 --> 00:13:38.560
I get the input methods documentation.

13:38.560 --> 00:13:40.639
I then delete other windows 

00:13:40.639 --> 00:13:45.440
and keep the help buffer visible.

13:45.440 --> 00:13:47.279
Notice that beh 

00:13:47.279 --> 00:13:52.720
in this input methods identifier.

13:52.720 --> 13:55.920
Here is the URL for full documentation

13:55.920 --> 00:14:00.639
on the web. The keyboard layout itself

00:14:00.639 --> 00:14:03.760
is a one-to-one mapping, but towards

00:14:03.760 --> 00:14:07.360
making transliteration intuitive.

14:07.360 --> 14:10.240
Multiple keys are sometimes mapped to

14:10.240 --> 14:12.079
the same letter.

14:12.079 --> 00:14:19.760
For example, both 'i' and 'y' produce yeh.

14:19.760 --> 14:22.959
The usual two letter transliterations

14:22.959 --> 00:14:29.600
ending with 'h' -- zh, ch, sh, and kh

00:14:29.600 --> 00:14:34.160
are provided. The ampersand prefix 

00:14:34.160 --> 00:14:37.440
is used to support often invisible

14:37.440 --> 00:14:41.600
BIDI markings. In addition to this

00:14:41.600 --> 00:14:43.920
internal documentation, 

00:14:43.920 --> 00:14:48.958
full documentation is also available.

14:48.959 --> 00:14:50.560
Complete documentation 

00:14:50.560 --> 00:14:52.240
for Persian input methods

00:14:52.240 --> 00:14:58.560
is available as PLPC-120036.

14:58.560 --> 00:15:00.880
Next, we'll take a quick look at this 

00:15:00.880 --> 00:15:03.600
on the web.

15:03.600 --> 00:15:05.440
You can click on links 

00:15:05.440 --> 00:15:07.440
in the reveal web-based form

00:15:07.440 --> 00:15:09.839
of this presentation.

15:09.839 --> 15:15.440
So let's visit PLPC-120036.

15:15.440 --> 00:15:17.600
This document fully describes 

00:15:17.600 --> 00:15:20.560
Persian input methods.

15:20.560 --> 00:15:22.320
In addition to HTML, 

00:15:22.320 --> 00:15:26.079
you can also obtain it in PDF. 

00:15:26.079 --> 00:15:30.079
Let's do that. Of particular interest 

00:15:30.079 --> 00:15:33.120
in this document are various tables

15:33.120 --> 00:15:35.839
that enumerate lists of letters 

00:15:35.839 --> 00:15:37.120
with their association 

00:15:37.120 --> 00:15:41.839
to both Persian input methods.

15:41.839 --> 15:46.240
Let's take a look at a few of these.

15:46.240 --> 00:15:51.360
Table 3, mapping of isiri-6219

15:51.360 --> 15:53.680
(the Farsi character set)

15:53.680 --> 00:15:55.680
to Emacs version input methods 

00:15:55.680 --> 00:15:58.720
could be of interest to you,

15:58.720 --> 16:01.040
as well as table 8

16:01.040 --> 16:04.720
for BIDI-related control markups,

16:04.720 --> 00:16:17.518
and table 9, for vowels and other signs.

16:17.519 --> 00:16:19.759
Having covered input methods, 

00:16:19.759 --> 00:16:21.279
let's turn our attention 

00:16:21.279 --> 00:16:25.199
to ramifications of BIDI and Perso-Arabic

00:16:25.199 --> 00:16:28.480
on various Emacs applications.

16:28.480 --> 00:16:32.160
Since 2012, I have been using Persian text

00:16:32.160 --> 00:16:35.839
in various Emacs applications.

16:35.839 --> 00:16:38.079
In short, my experience has been 

00:16:38.079 --> 00:16:42.399
that most Emacs apps are usable,

16:42.399 --> 16:44.880
but they all have glitches

16:44.880 --> 00:16:47.279
that could at a minimum 

00:16:47.279 --> 00:16:51.279
annoy Perso-Arabic users.

16:51.279 --> 16:54.639
In this slide, I'm presenting a summary.

16:54.639 --> 00:16:56.399
The glitches with Gnus 

00:16:56.399 --> 00:17:00.079
are not all that significant for me.

17:00.079 --> 17:04.240
BBDB glitches can easily be fixed.

17:04.240 --> 00:17:07.679
For Calendar, I have customized my own setup

00:17:07.679 --> 00:17:10.799
to support Persian and Islamic dates

00:17:10.799 --> 00:17:14.079
in Perso-Arabic. Perhaps they should be

00:17:14.079 --> 00:17:17.039
merged upstream, instead of 

00:17:17.039 --> 00:17:19.679
dealing with apps one at a time.

00:17:19.679 --> 00:17:21.439
I think it's more reasonable 

00:17:21.439 --> 00:17:25.279
to consider them collectively.

17:25.280 --> 00:17:26.880
The glitches that I mentioned 

00:17:26.880 --> 00:17:28.799
in the previous slide

17:28.799 --> 17:30.480
have two routes:

17:30.480 --> 17:32.480
some are BIDI-specific

17:32.480 --> 17:35.919
and some are Perso-Arabic-specific.

17:35.919 --> 00:17:36.720
In this slide, 

00:17:36.720 --> 00:17:39.280
I have classified them as such

17:39.280 --> 17:42.960
and have made some general suggestions,

17:42.960 --> 00:17:45.679
but all of these at best amount to

00:17:45.679 --> 00:17:47.679
tactical approaches.

17:47.679 --> 00:17:50.160
I think a more strategic approach 

00:17:50.160 --> 00:17:53.519
is called for.

17:53.520 --> 17:56.240
The right way to address BIDI-awareness

17:56.240 --> 17:58.559
and other awarenesses

17:58.559 --> 00:18:00.559
is to build them in frameworks 

00:18:00.559 --> 00:18:04.160
that Emacs apps can then use.

18:04.160 --> 00:18:08.480
So, I'm proposing that we first create ENML, 

00:18:08.480 --> 00:18:11.520
the Emacs Native Markup Language, 

00:18:11.520 --> 00:18:16.880
as a Lisp-ish (perhaps even not fully secure)

18:16.880 --> 18:20.320
super-set of HTML5.

18:20.320 --> 18:22.960
With that in place, we can then build on

18:22.960 --> 00:18:25.360
the two decades of experience 

00:18:25.360 --> 00:18:27.600
that have produced various 

00:18:27.600 --> 00:18:31.039
web application development frameworks

18:31.039 --> 18:33.919
by mimicking one of them.

18:33.919 --> 00:18:35.679
I don't have any running code 

00:18:35.679 --> 00:18:39.440
for any of these, but discussing strategy 

00:18:39.440 --> 00:18:44.639
need not always be futile.

18:44.640 --> 18:47.679
Emacs has immense potentials,

18:47.679 --> 18:51.120
but those potentials cannot be realized

18:51.120 --> 00:18:54.640
unless we integrate Emacs in the totality

00:18:54.640 --> 00:18:59.840
of a specific complete digital ecosystem.

18:59.840 --> 19:02.000
Over the past two decades, I've been

19:02.000 --> 00:19:03.440
building the contours 

00:19:03.440 --> 00:19:06.240
of The Libre-Halaal By* (ByStar) 

00:19:06.240 --> 00:19:10.240
digital ecosystem.

19:10.240 --> 19:13.039
Emacs can then be fully integrated into

19:13.039 --> 00:19:16.799
ByStar. It's through such integration 

00:19:16.799 --> 00:19:19.679
that full conviviality of Emacs

00:19:19.679 --> 00:19:24.239
can be experienced.

19:24.240 --> 00:19:26.640
Blee, the ByStar Libre-Halaal 

00:19:26.640 --> 00:19:30.000
Emacs environment is Emacs 

00:19:30.000 --> 00:19:32.720
plus a whole lot of Emacs apps

00:19:32.720 --> 00:19:34.880
integrated with Debian,

00:19:34.880 --> 00:19:38.320
With ByStar services and with BISOS, 

00:19:38.320 --> 00:19:42.480
the ByStar Internet Services OS.

19:42.480 --> 00:19:43.600
Perhaps this could be 

00:19:43.600 --> 00:19:45.840
the topic of a presentation

00:19:45.840 --> 00:19:46.840
for the 2022 Emacs Conference.

00:19:46.840 --> 00:19:51.559
[captions by Mohsen Banan]