summaryrefslogtreecommitdiffstats
diff options
context:
space:
mode:
Diffstat (limited to '')
-rw-r--r--2021/captions/emacsconf-2021-structural--tree-edit-structural-editing-for-java-python-c-and-beyond--ethan-leba--main.vtt817
-rw-r--r--2021/info/faster-schedule.md2
-rw-r--r--2021/info/structural-schedule.md10
-rw-r--r--2021/schedule-details.md4
4 files changed, 827 insertions, 6 deletions
diff --git a/2021/captions/emacsconf-2021-structural--tree-edit-structural-editing-for-java-python-c-and-beyond--ethan-leba--main.vtt b/2021/captions/emacsconf-2021-structural--tree-edit-structural-editing-for-java-python-c-and-beyond--ethan-leba--main.vtt
new file mode 100644
index 00000000..9bba8336
--- /dev/null
+++ b/2021/captions/emacsconf-2021-structural--tree-edit-structural-editing-for-java-python-c-and-beyond--ethan-leba--main.vtt
@@ -0,0 +1,817 @@
+WEBVTT
+
+00:00.080 --> 00:01.360
+Hi. My name is Ethan,
+
+00:01.360 --> 00:02.320
+and today I'm going to be speaking
+
+00:02.320 --> 00:04.240
+about tree-edit, which is a package
+
+00:04.240 --> 00:06.160
+which aims to bring structural editing
+
+00:06.160 --> 00:08.320
+to everyday languages.
+
+00:08.320 --> 00:10.559
+So what is structural editing?
+
+00:10.559 --> 00:11.657
+The way that we typically
+
+00:11.657 --> 00:12.578
+write code today
+
+00:12.578 --> 00:14.480
+is working with characters, words,
+
+00:14.480 --> 00:16.206
+lines, paragraphs, and so on,
+
+00:16.206 --> 00:18.600
+and these objects have no real relation
+
+00:18.600 --> 00:21.520
+to the structure of programming languages.
+
+00:21.520 --> 00:24.667
+In contrast, tree-edit's editing operations
+
+00:24.667 --> 00:26.897
+map exactly to the structure
+
+00:26.897 --> 00:28.411
+of the programming language,
+
+00:28.411 --> 00:30.303
+which is typically in a tree form
+
+00:30.303 --> 00:32.053
+with different types of nodes
+
+00:32.053 --> 00:33.920
+such as identifiers, expressions,
+
+00:33.920 --> 00:35.957
+and statements. Using this structure
+
+00:35.957 --> 00:37.548
+can enable much more powerful
+
+00:37.548 --> 00:39.200
+editing operations,
+
+00:39.200 --> 00:40.769
+and crucially editing operations
+
+00:40.769 --> 00:42.081
+that map much more closely
+
+00:42.081 --> 00:44.960
+to the way that we think about code.
+
+00:44.960 --> 00:46.140
+tree-edit was inspired by
+
+00:46.140 --> 00:47.386
+paredit and lispy,
+
+00:47.386 --> 00:48.320
+which are two great
+
+00:48.320 --> 00:50.271
+Lisp structural editors.
+
+00:50.271 --> 00:52.383
+However, what makes tree-edit unique
+
+00:52.383 --> 00:54.480
+is that it can work with many languages,
+
+00:54.480 --> 00:55.759
+such as some of the
+
+00:55.759 --> 00:59.826
+more mainstream languages like C, Java,
+
+00:59.826 --> 01:01.600
+Python, and so on.
+
+01:01.600 --> 01:03.273
+So now I'm going to show off tree-edit
+
+01:03.273 --> 01:05.705
+in action, working with a Java program.
+
+01:05.705 --> 01:07.237
+So we can see on the left,
+
+01:07.237 --> 01:09.119
+we have a syntax tree,
+
+01:09.119 --> 01:11.560
+and the node in bold is what I call
+
+01:11.560 --> 01:13.780
+the current node. So instead of
+
+01:13.780 --> 01:15.100
+the concept of a cursor,
+
+01:15.100 --> 01:17.600
+where we have a point in 2D space,
+
+01:17.600 --> 01:20.285
+we instead work with a current node
+
+01:20.285 --> 01:22.729
+which all our editing operations
+
+01:22.729 --> 01:23.840
+take place upon.
+
+01:23.840 --> 01:26.479
+So we can move up and down,
+
+01:26.479 --> 01:28.720
+or rather side to side,
+
+01:28.720 --> 01:31.160
+move inwards down to the children
+
+01:31.160 --> 01:33.920
+of the tree, back up to the parents.
+
+01:33.920 --> 01:36.799
+We can also jump to a node by its type.
+
+01:36.799 --> 01:38.768
+So we're going to jump to
+
+01:38.768 --> 01:40.880
+a variable declaration.
+
+01:40.880 --> 01:44.399
+We can jump to an if statement.
+
+01:44.399 --> 01:46.880
+And as you might have noticed,
+
+01:46.880 --> 01:48.360
+tree-edit by default
+
+01:48.360 --> 01:51.337
+uses a vim-style mode of editing,
+
+01:51.337 --> 01:55.119
+so it's a verb, which would be jump,
+
+01:55.119 --> 01:56.874
+and then a type,
+
+01:56.874 --> 02:00.799
+which would be if statement.
+
+02:00.799 --> 02:03.346
+So now I'll show off
+
+02:03.346 --> 02:06.144
+the syntax tree modification in action.
+
+02:06.144 --> 02:08.000
+So if I delete this deleteme node,
+
+02:08.000 --> 02:10.112
+we can see the node is deleted,
+
+02:10.112 --> 02:12.049
+and also the comma is removed
+
+02:12.049 --> 02:13.920
+since it's no longer needed.
+
+02:13.920 --> 02:16.720
+We can add some nodes back in.
+
+02:16.720 --> 02:18.160
+Here we just have a placeholder node
+
+02:18.160 --> 02:20.391
+called tree, which we can swap out
+
+02:20.391 --> 02:21.875
+with whatever we like.
+
+02:21.875 --> 02:24.560
+So if we want to put in, for example,
+
+02:24.560 --> 02:29.280
+a plus or minus operator,
+
+02:29.280 --> 02:30.879
+it'll put these two TREE things here
+
+02:30.879 --> 02:32.634
+since there needs to be something there,
+
+02:32.634 --> 02:37.360
+but we can go fill them out as we like.
+
+02:37.360 --> 02:38.595
+So that's what that is.
+
+02:38.595 --> 02:41.920
+Then I'll delete these again.
+
+02:41.920 --> 02:43.709
+Next we can see raising.
+
+02:43.709 --> 02:45.280
+So if I raise reader,
+
+02:45.280 --> 02:46.160
+then it will replace
+
+02:46.160 --> 02:47.342
+the outer function call
+
+02:47.342 --> 02:48.583
+with the node itself.
+
+02:48.583 --> 02:50.948
+I could raise it again.
+
+02:50.948 --> 02:53.363
+The opposite operation to that
+
+02:53.363 --> 02:57.200
+is wrapping. So I can wrap reader
+
+02:57.200 --> 02:59.519
+back into function call,
+
+02:59.519 --> 03:03.009
+and I could wrap this again
+
+03:03.009 --> 03:08.480
+if I wanted to. So that is wrapping.
+
+03:08.480 --> 03:12.640
+We can also do it on a statement level,
+
+03:12.640 --> 03:13.760
+so if I want to wrap this
+
+03:13.760 --> 03:14.480
+in an if statement,
+
+03:14.480 --> 03:17.034
+I can wrap the statement,
+
+03:17.034 --> 03:18.400
+and there we go.
+
+03:18.400 --> 03:21.280
+And let's just raise it back up,
+
+03:21.280 --> 03:23.200
+raise it again.
+
+03:23.200 --> 03:25.760
+There we go. Finally, I'll show off
+
+03:26.959 --> 03:28.720
+slurping and barfing,
+
+03:28.720 --> 03:32.256
+which... a little bit gross words,
+
+03:32.256 --> 03:34.879
+but I think it accurately describes
+
+03:34.879 --> 03:37.519
+the action, so let me just add
+
+03:37.519 --> 03:41.120
+a couple breaks here.
+
+03:41.120 --> 03:44.748
+So let's say we want
+
+03:44.748 --> 03:46.779
+this if statement and a couple of breaks
+
+03:46.779 --> 03:48.319
+to be inside of the while,
+
+03:48.319 --> 03:50.959
+so we can just slurp this up,
+
+03:50.959 --> 03:52.433
+and if we don't actually want them,
+
+03:52.433 --> 03:54.528
+we can barf them back out.
+
+03:54.528 --> 03:56.736
+So that's where those words
+
+03:56.736 --> 03:57.840
+have come from.
+
+03:57.840 --> 04:01.120
+And we can just... delete as we please.
+
+04:01.120 --> 04:03.826
+So yeah, that's a quick overview
+
+04:03.826 --> 04:07.360
+of the tree editing plugin in action.
+
+04:07.360 --> 04:08.900
+So now I want to talk a little bit
+
+04:08.900 --> 04:12.080
+about the implementation of tree-edit.
+
+04:12.080 --> 04:14.400
+Tree-edit uses the tree-sitter parser
+
+04:14.400 --> 04:17.919
+to convert text into a syntax tree.
+
+04:17.919 --> 04:21.501
+Tree-sitter is used by GitHub
+
+04:21.501 --> 04:22.752
+for its syntax highlighting,
+
+04:22.752 --> 04:25.280
+and it's available in a bunch of editors,
+
+04:25.280 --> 04:27.120
+including Emacs, so it's
+
+04:27.120 --> 04:28.960
+a fairly standard tool.
+
+04:28.960 --> 04:30.960
+However, the unique part about tree-edit
+
+04:30.960 --> 04:32.479
+is how it performs
+
+04:32.479 --> 04:34.479
+correct editing operations
+
+04:34.479 --> 04:35.919
+on the syntax tree
+
+04:35.919 --> 04:38.320
+and then converts that back into text.
+
+04:38.320 --> 04:41.759
+So to do that, we use miniKanren,
+
+04:41.759 --> 04:43.759
+and miniKanren is an embedded
+
+04:43.759 --> 04:45.120
+domain-specific language
+
+04:45.120 --> 04:47.440
+for logic programming.
+
+04:47.440 --> 04:50.080
+So what exactly does that mean?
+
+04:50.080 --> 04:51.280
+In our case, it's just
+
+04:51.280 --> 04:54.240
+an Emacs Lisp library called reazon,
+
+04:54.240 --> 04:56.720
+which exposes a set of macros
+
+04:56.720 --> 04:58.320
+which enables us to program
+
+04:58.320 --> 05:01.360
+in this logic programming style.
+
+05:01.360 --> 05:03.280
+I'm not going to get into the details
+
+05:03.280 --> 05:05.520
+of how logic programming works.
+
+05:05.520 --> 05:07.520
+However, one of the most unique aspects
+
+05:07.520 --> 05:09.919
+about it is that we can define
+
+05:09.919 --> 05:13.600
+a predicate and then figure out
+
+05:13.600 --> 05:15.280
+all the inputs to the predicate
+
+05:15.280 --> 05:17.759
+that would hold to be true.
+
+05:17.759 --> 05:19.360
+So in this case,
+
+05:19.360 --> 05:21.520
+we have our query variable q,
+
+05:21.520 --> 05:24.479
+which will be what the output is,
+
+05:24.479 --> 05:29.120
+and we are asking for all the values of q
+
+05:29.120 --> 05:32.080
+that pass this predicate of
+
+05:32.080 --> 05:34.479
+being set-equal to 1 2 3 4.
+
+05:34.479 --> 05:36.880
+So if we execute this,
+
+05:36.880 --> 05:40.080
+it will take a little time...
+
+05:40.080 --> 05:41.520
+It shouldn't be taking this long.
+
+05:41.520 --> 05:43.280
+Oh, there it goes.
+
+05:43.280 --> 05:45.919
+We can see that it's generated
+
+05:45.919 --> 05:47.520
+a bunch of different answers
+
+05:47.520 --> 05:51.199
+that are all set-equal to 1 2 3 4.
+
+05:51.199 --> 05:52.880
+So it's just a bunch of
+
+05:52.880 --> 05:57.280
+different permutations of that.
+
+05:57.280 --> 05:59.120
+We can extend this notion
+
+05:59.120 --> 06:03.600
+to a parser. In tree-edit, we've defined
+
+06:03.600 --> 06:05.360
+a parser in reazon,
+
+06:05.360 --> 06:10.800
+and we can use that parser to figure out
+
+06:10.800 --> 06:15.919
+any tokens that match the type of node
+
+06:15.919 --> 06:16.880
+that we're trying to generate.
+
+06:16.880 --> 06:19.600
+If I execute this, we can see
+
+06:19.600 --> 06:21.199
+that reazon has generated
+
+06:21.199 --> 06:23.440
+these five answers that match
+
+06:23.440 --> 06:26.960
+what a try statement is in Java.
+
+06:26.960 --> 06:29.680
+Here we can see we can have
+
+06:29.680 --> 06:31.919
+an infinite amount of catches
+
+06:31.919 --> 06:34.720
+optionally ending with a finally,
+
+06:34.720 --> 06:36.160
+and we always have to start
+
+06:36.160 --> 06:39.039
+with a try and a block.
+
+06:39.039 --> 06:40.000
+We can see this again
+
+06:40.000 --> 06:42.400
+with an argument list.
+
+06:42.400 --> 06:43.520
+We have the opening and closing
+
+06:43.520 --> 06:45.759
+parentheses, and expressions
+
+06:45.759 --> 06:49.120
+which are comma delimited.
+
+06:49.120 --> 06:51.759
+Now, for a more complex example, and
+
+06:51.759 --> 06:53.680
+something that is along the lines
+
+06:53.680 --> 06:55.199
+of what's in tree-edit,
+
+06:55.199 --> 06:57.919
+is if we have this x here
+
+06:57.919 --> 07:01.599
+and we want to insert another expression,
+
+07:01.599 --> 07:05.759
+so x, y. We can assert
+
+07:05.759 --> 07:07.680
+that there's some new tokens,
+
+07:07.680 --> 07:10.160
+and we want an expression
+
+07:10.160 --> 07:11.840
+to be in those new tokens,
+
+07:11.840 --> 07:13.280
+and we can essentially state
+
+07:13.280 --> 07:15.039
+where we want these new tokens to go
+
+07:15.039 --> 07:19.759
+within the old list of tokens,
+
+07:19.759 --> 07:21.599
+so replacing it
+
+07:21.599 --> 07:23.360
+after the previous expression,
+
+07:23.360 --> 07:26.000
+before the closed parentheses,
+
+07:26.000 --> 07:26.880
+and then we can state
+
+07:26.880 --> 07:28.560
+that the whole thing parses.
+
+07:28.560 --> 07:30.080
+If we run that, we can see that
+
+07:30.080 --> 07:32.479
+as we wanted earlier,
+
+07:32.479 --> 07:37.120
+which was a comma and then expression,
+
+07:37.120 --> 07:39.120
+we have that here as well.
+
+07:39.120 --> 07:41.759
+We can see this again.
+
+07:41.759 --> 07:42.720
+Here, the only change is that
+
+07:42.720 --> 07:45.280
+we've moved the tokens to be
+
+07:45.280 --> 07:46.240
+before the expression.
+
+07:46.240 --> 07:48.800
+So we want to put an expression
+
+07:48.800 --> 07:50.560
+before this x, so we want something
+
+07:50.560 --> 07:52.560
+like y, x,
+
+07:52.560 --> 07:54.240
+and if we execute that,
+
+07:54.240 --> 07:57.919
+we can see that it is correctly asserted
+
+07:57.919 --> 07:59.039
+that it would be an expression
+
+07:59.039 --> 08:01.520
+and then a comma afterwards.
+
+08:01.520 --> 08:02.960
+One last example is
+
+08:02.960 --> 08:04.400
+if we have an if statement
+
+08:04.400 --> 08:07.759
+and we want to add an extra block,
+
+08:07.759 --> 08:11.599
+we can see that it correctly figures out
+
+08:11.599 --> 08:12.400
+that we need an else
+
+08:12.400 --> 08:13.840
+in order to have another statement
+
+08:13.840 --> 08:16.720
+in an if statement.
+
+08:16.720 --> 08:19.759
+So, next steps for tree-edit.
+
+08:19.759 --> 08:21.039
+The core of tree-edit is in place
+
+08:21.039 --> 08:23.120
+but there's a lot of usability features
+
+08:23.120 --> 08:25.360
+to add, and a lot of testing
+
+08:25.360 --> 08:26.400
+that needs to be done
+
+08:26.400 --> 08:29.599
+in order to iron out any bugs that exist.
+
+08:29.599 --> 08:30.960
+I'd like to add support
+
+08:30.960 --> 08:35.200
+for as many languages as is possible.
+
+08:35.200 --> 08:36.240
+I think my next step
+
+08:36.240 --> 08:38.490
+will probably be Python.
+
+08:38.490 --> 08:41.279
+There's some performance improvements
+
+08:41.279 --> 08:44.080
+that need to be made, since using this
+
+08:44.080 --> 08:45.519
+logic programming language
+
+08:45.519 --> 08:47.600
+is fairly intensive.
+
+08:47.600 --> 08:48.800
+There's some optimizations
+
+08:48.800 --> 08:50.560
+both on the library side
+
+08:50.560 --> 08:51.519
+and on tree-edit side
+
+08:51.519 --> 08:53.360
+that can be made.
+
+08:53.360 --> 08:55.519
+Contributors are of course welcome,
+
+08:55.519 --> 09:00.000
+as tree-edit is an open source project.
+
+09:00.000 --> 09:03.360
+For future work, I think the prospect
+
+09:03.360 --> 09:04.480
+of voice controlled development
+
+09:04.480 --> 09:06.240
+with tree-edit is actually something
+
+09:06.240 --> 09:07.920
+that's really exciting,
+
+09:07.920 --> 09:11.120
+since syntax can be very cumbersome
+
+09:11.120 --> 09:12.320
+when you're working with
+
+09:12.320 --> 09:14.240
+voice control software.
+
+09:14.240 --> 09:16.320
+I can envision something like
+
+09:16.320 --> 09:19.440
+saying, "Jump to identifier,
+
+09:19.440 --> 09:26.640
+add plus operator, jump to if statement,
+
+09:26.640 --> 09:30.480
+wrap if statement in while."
+
+09:30.480 --> 09:31.519
+So that's something
+
+09:31.519 --> 09:33.519
+I'd like to investigate.
+
+09:33.519 --> 09:35.040
+I also would just like to
+
+09:35.040 --> 09:37.279
+provide the core functionality
+
+09:37.279 --> 09:39.120
+of [tree-edit] as something
+
+09:39.120 --> 09:40.399
+that can be used as a library
+
+09:40.399 --> 09:41.920
+for other projects,
+
+09:41.920 --> 09:43.839
+such as refactoring packages,
+
+09:43.839 --> 09:46.240
+or other non-Vim-style approaches,
+
+09:46.240 --> 09:49.200
+and just making the syntax generation
+
+09:49.200 --> 09:52.080
+available for reuse.
+
+09:52.080 --> 09:53.760
+Finally, I'd like to thank
+
+09:53.760 --> 09:56.399
+the authors of reazon
+
+09:56.399 --> 09:58.399
+and elisp-tree-sitter,
+
+09:58.399 --> 10:00.185
+which in turn packages
+
+10:00.185 --> 10:02.079
+tree-sitter itself,
+
+10:02.079 --> 10:05.440
+since tree-edit relies very heavily
+
+10:05.440 --> 10:07.680
+on these two packages.
+
+10:07.680 --> 10:08.959
+I'd also like to thank
+
+10:08.959 --> 10:10.480
+the author of lispy,
+
+10:10.480 --> 10:12.720
+since a lot of the design decisions
+
+10:12.720 --> 10:14.800
+when it comes to the editing operations
+
+10:14.800 --> 10:18.560
+are based very heavily on lispy.
+
+10:18.560 --> 10:20.320
+So that's the end of my talk.
+
+10:20.320 --> 10:22.959
+Thank you for watching.
+
+10:22.959 --> 10:23.959
+[captions by sachac]
diff --git a/2021/info/faster-schedule.md b/2021/info/faster-schedule.md
index 60af8009..9374d079 100644
--- a/2021/info/faster-schedule.md
+++ b/2021/info/faster-schedule.md
@@ -1,7 +1,7 @@
<!-- Automatically generated by conf-create-info-pages -->
Q&A: live
-Status: Now playing
+Status: Finished
Duration: 35:35
diff --git a/2021/info/structural-schedule.md b/2021/info/structural-schedule.md
index d226e3c5..93ad29e0 100644
--- a/2021/info/structural-schedule.md
+++ b/2021/info/structural-schedule.md
@@ -1,10 +1,14 @@
<!-- Automatically generated by conf-create-info-pages -->
Q&A: live
-Status: Captions added to video
+Status: Now playing
Duration: 10:24
-<div class="times" start="2021-11-28T14:58:00Z" end="2021-11-28T15:09:00Z">Sunday, Nov 28 2021, ~ 9:58 AM - 10:09 AM EST<br />Sunday, Nov 28 2021, ~ 6:58 AM - 7:09 AM PST<br />Sunday, Nov 28 2021, ~ 2:58 PM - 3:09 PM UTC<br />Sunday, Nov 28 2021, ~ 3:58 PM - 4:09 PM CET<br />Sunday, Nov 28 2021, ~ 4:58 PM - 5:09 PM EET<br />Sunday, Nov 28 2021, ~ 8:28 PM - 8:39 PM IST<br />Sunday, Nov 28 2021, ~10:58 PM - 11:09 PM +08<br />Sunday, Nov 28 2021, ~11:58 PM - 12:09 AM JST<br /><a href="/2021/">Find out how to watch and participate</a></div>
-# Description
+
+<div class="mainVideo"><div class="video-card vid" data-id="mainVideo"><figure><video controls preload="metadata" id="mainVideo">
+<source src="https://media.emacsconf.org/2021/emacsconf-2021-structural--tree-edit-structural-editing-for-java-python-c-and-beyond--ethan-leba--main.webm"><track label="English" kind="captions" srclang="en" src="/2021/captions/emacsconf-2021-structural--tree-edit-structural-editing-for-java-python-c-and-beyond--ethan-leba--main.vtt" default>
+</video></figure>
+<div class="files resources"><ul><li><a href="https://media.emacsconf.org/2021/emacsconf-2021-structural--tree-edit-structural-editing-for-java-python-c-and-beyond--ethan-leba--main.webm">Download .webm video (10:24, 13MB)</a></li><li><a href="https://media.emacsconf.org/2021/emacsconf-2021-structural--tree-edit-structural-editing-for-java-python-c-and-beyond--ethan-leba--main.vtt">Download --main.vtt</a></li></ul></div></div>
+</div># Description
diff --git a/2021/schedule-details.md b/2021/schedule-details.md
index e0f75ff5..5a534fc1 100644
--- a/2021/schedule-details.md
+++ b/2021/schedule-details.md
@@ -35,8 +35,8 @@
<tr><td colspan="4"><strong>Sunday, Nov 28, 2021: Development talks<strong></td></tr>
<tr><td colspan="4"><strong>Sunday morning<strong></td></tr>
<tr><td></td><td width=100>~ 9:00 AM</td><td><a href="/2021/talks/day2-open">Opening remarks day 2</a></td><td></td></tr>
-<tr><td>now playing</td><td width=100>~ 9:07 AM</td><td><a href="/2021/talks/faster">Optimizing Emacs Lisp Code</a></td><td>Dmitry Gutov</td></tr>
-<tr><td>captioned</td><td width=100>~ 9:58 AM</td><td><a href="/2021/talks/structural">Tree-edit: Structural editing for Java, Python, C, and beyond!</a></td><td>Ethan Leba</td></tr>
+<tr><td>done</td><td width=100>~ 9:07 AM</td><td><a href="/2021/talks/faster">Optimizing Emacs Lisp Code</a></td><td>Dmitry Gutov</td></tr>
+<tr><td>now playing</td><td width=100>~10:00 AM</td><td><a href="/2021/talks/structural">Tree-edit: Structural editing for Java, Python, C, and beyond!</a></td><td>Ethan Leba</td></tr>
<tr><td>captioned</td><td width=100>~10:15 AM</td><td><a href="/2021/talks/ui">Yak-shaving to a UI framework</a></td><td>Erik Anderson</td></tr>
<tr><td>captioned</td><td width=100>~10:35 AM</td><td><a href="/2021/talks/mold">Moldable Emacs, a step towards sustainable software</a></td><td>Andrea</td></tr>
<tr><td>captioned</td><td width=100>~10:49 AM</td><td><a href="/2021/talks/model">Extending the "model" of Emacs to other applications</a></td><td>Laszlo Krajnikovszkij</td></tr>