1
00:00:00,000 --> 00:00:01,083
DAVID MALAN: Hello, world.

2
00:00:01,083 --> 00:00:04,620
My name is David Malan, and I'm here in
Sanders Theatre at Harvard University

3
00:00:04,620 --> 00:00:06,720
where I teach CS50,
Harvard's introduction

4
00:00:06,720 --> 00:00:08,940
to the intellectual
enterprises of computer science

5
00:00:08,940 --> 00:00:11,940
and the arts of programming, an
introductory course for majors and non

6
00:00:11,940 --> 00:00:13,020
majors alike.

7
00:00:13,020 --> 00:00:16,010
CS50 is among Harvard's largest
courses, and indeed in healthier times,

8
00:00:16,010 --> 00:00:18,510
the seats behind me would look
a little something like this.

9
00:00:18,510 --> 00:00:21,030
But the course is also freely
available as OpenCourseWare

10
00:00:21,030 --> 00:00:23,970
as CS50x via platforms like
YouTube, and edX, and beyond.

11
00:00:23,970 --> 00:00:26,970
And the course is additionally available
to teachers in the high schools

12
00:00:26,970 --> 00:00:29,520
and middle schools as
CS50 AP, a curriculum

13
00:00:29,520 --> 00:00:32,130
that they can adopt or adapt
for their own classroom.

14
00:00:32,130 --> 00:00:34,800
Students meanwhile can take
both before and after CS50

15
00:00:34,800 --> 00:00:38,050
itself, a whole suite of
courses nowadays by CS50's team,

16
00:00:38,050 --> 00:00:41,010
including courses on artificial
intelligence, game development, web

17
00:00:41,010 --> 00:00:42,990
programming, and more.

18
00:00:42,990 --> 00:00:47,760
All of these courses meanwhile are
available at edx.org/cs50 for free.

19
00:00:47,760 --> 00:00:50,190
Now, over the years have
we had to provide students

20
00:00:50,190 --> 00:00:51,982
with quite a bit of
infrastructure in order

21
00:00:51,982 --> 00:00:54,440
to make this and now other
courses possible, particularly

22
00:00:54,440 --> 00:00:56,940
toward an end to onboarding
them to the world of programming

23
00:00:56,940 --> 00:01:00,420
without having to configure and install
software on their own Macs and PCs,

24
00:01:00,420 --> 00:01:03,000
invariably then having
technical support difficulties

25
00:01:03,000 --> 00:01:06,930
as opposed to focusing on the course's
actual curriculum and the principles

26
00:01:06,930 --> 00:01:08,670
of programming itself at term starts.

27
00:01:08,670 --> 00:01:11,730
And in fact, if we roll back to
2007 at the time with students

28
00:01:11,730 --> 00:01:15,060
here only on campus being provided
with shell accounts on a Unix,

29
00:01:15,060 --> 00:01:18,180
later, Linux cluster on campus for
which they had a username and password.

30
00:01:18,180 --> 00:01:22,673
They had a home directory as well as
access to commands like GCC and GDB

31
00:01:22,673 --> 00:01:25,590
because indeed the course introduces
students to programming initially

32
00:01:25,590 --> 00:01:28,990
by way of C, followed by Python
and other languages as well.

33
00:01:28,990 --> 00:01:31,950
But in 2008, we began to
experiment with the cloud,

34
00:01:31,950 --> 00:01:33,540
which was in its earliest form.

35
00:01:33,540 --> 00:01:35,970
And by using AWS EC2
and other services did

36
00:01:35,970 --> 00:01:38,430
we replicate that same
topology in the cloud so as

37
00:01:38,430 --> 00:01:40,740
to actually have more
administrative and educational

38
00:01:40,740 --> 00:01:45,840
control over that same environment for
just our students in the course itself.

39
00:01:45,840 --> 00:01:48,450
In 2012 meanwhile,
just a few years later,

40
00:01:48,450 --> 00:01:51,090
did we begin to experiment with
downloadable virtual machines,

41
00:01:51,090 --> 00:01:54,918
in our cases, dubbed the CS50 Appliance,
a Fedora and later Ubuntu image

42
00:01:54,918 --> 00:01:57,210
that students could download
on their own Macs and PCs.

43
00:01:57,210 --> 00:02:00,000
They could download something
like VirtualBox or VMware

44
00:02:00,000 --> 00:02:03,630
in order to run their own CS50
environment on their own computer.

45
00:02:03,630 --> 00:02:06,690
Particularly enabling was this
for the OpenCourseWare audience

46
00:02:06,690 --> 00:02:09,755
who, to date, could watch content,
could read PDFs and the like,

47
00:02:09,755 --> 00:02:12,630
but they couldn't necessarily do
the course's programming assignments

48
00:02:12,630 --> 00:02:16,260
or problem sets not having access of
course to Harvard's shell accounts here

49
00:02:16,260 --> 00:02:16,980
on campus.

50
00:02:16,980 --> 00:02:20,188
But with the Appliance were they able
to work not only on their own computers

51
00:02:20,188 --> 00:02:25,470
but even offline, engaging actively
in the course's OpenCourseWare.

52
00:02:25,470 --> 00:02:29,520
But a few years later in 2015 did we
begin to explore web-based alternatives

53
00:02:29,520 --> 00:02:32,520
as well to just get away from
the process of needing students

54
00:02:32,520 --> 00:02:34,650
to install anything
on their Macs or PCs,

55
00:02:34,650 --> 00:02:38,010
particularly in the context of schools,
where teachers might themselves not

56
00:02:38,010 --> 00:02:40,560
have administrative access
over the school's own labs.

57
00:02:40,560 --> 00:02:43,290
And so we began to experiment
with an open source tool

58
00:02:43,290 --> 00:02:46,890
at the time, now AWS Cloud9,
on top of which we developed

59
00:02:46,890 --> 00:02:50,220
a suite of pedagogical plug-ins so
that students could use just a browser

60
00:02:50,220 --> 00:02:53,100
to access still a text
editor, and a terminal window,

61
00:02:53,100 --> 00:02:56,610
and all of those same tools but
this time requiring only a browser.

62
00:02:56,610 --> 00:03:00,120
And most recently now in 2021 have
we begun to experiment as well

63
00:03:00,120 --> 00:03:02,550
with Visual Studio Code
and its suite of extensions

64
00:03:02,550 --> 00:03:04,090
replicating that same environment.

65
00:03:04,090 --> 00:03:07,710
But we're also now aspiring to provide
students with an offboarding experience

66
00:03:07,710 --> 00:03:11,190
as well, whereby students at terms
end will now be able to download--

67
00:03:11,190 --> 00:03:14,490
once they are comfortable and have
their footing in programming itself--

68
00:03:14,490 --> 00:03:16,920
VS Code and any requisite
extensions in order

69
00:03:16,920 --> 00:03:19,170
to have their own independent
programming environment.

70
00:03:19,170 --> 00:03:21,940
Because indeed if CS50 is the
only proper programming class

71
00:03:21,940 --> 00:03:23,940
they ultimately take,
they can at least continue

72
00:03:23,940 --> 00:03:27,670
developing even without any of the
course's infrastructure thereafter.

73
00:03:27,670 --> 00:03:29,598
So in order to achieve
all of these tools,

74
00:03:29,598 --> 00:03:32,640
we were motivated by trying to solve
just these three problems-- enabling

75
00:03:32,640 --> 00:03:34,620
students to write
code, enabling students

76
00:03:34,620 --> 00:03:39,030
to submit code, and enable teachers
like ourselves to review that same code.

77
00:03:39,030 --> 00:03:41,400
And underlying then
all of those problems

78
00:03:41,400 --> 00:03:44,490
were a number of problems already
solved, for instance, authentication,

79
00:03:44,490 --> 00:03:47,820
authorization, code reviews,
container orchestration, testing,

80
00:03:47,820 --> 00:03:49,420
version control, and more.

81
00:03:49,420 --> 00:03:52,560
And even though we have
had for years, myself

82
00:03:52,560 --> 00:03:56,070
included, this tendency
to try to create solutions

83
00:03:56,070 --> 00:03:59,010
ourselves and implement our
own homegrown solutions so as

84
00:03:59,010 --> 00:04:01,618
to get these tools just right,
that's of course intention

85
00:04:01,618 --> 00:04:04,410
with using something off the shelf,
which, while already developed,

86
00:04:04,410 --> 00:04:06,700
might not do exactly what you want.

87
00:04:06,700 --> 00:04:09,780
But we've over time tried to
find the sweet spot, whereby

88
00:04:09,780 --> 00:04:13,800
we use as much open source software
or SaaS type tools as we can

89
00:04:13,800 --> 00:04:16,829
and then stitch them together using
our own extensions, and plug-ins,

90
00:04:16,829 --> 00:04:20,422
and the like to create, ultimately, the
UX that we want our students to have.

91
00:04:20,422 --> 00:04:22,630
And so in the case of these
primitives, for instance,

92
00:04:22,630 --> 00:04:25,770
have we in recent years leverage GitHub
itself for all of these and more.

93
00:04:25,770 --> 00:04:29,100
For authentication, we have students
log in via OAuth with their own GitHub

94
00:04:29,100 --> 00:04:30,332
usernames and passwords.

95
00:04:30,332 --> 00:04:32,040
For authorization,
we've leveraged GitHub

96
00:04:32,040 --> 00:04:34,890
support for teams so that TAs
and students can have read

97
00:04:34,890 --> 00:04:36,840
and/or write access to the same repos.

98
00:04:36,840 --> 00:04:40,260
For code reviews, we've leveraged the
web UI's interface for pull requests

99
00:04:40,260 --> 00:04:43,440
or commits on which you
can type per line comments.

100
00:04:43,440 --> 00:04:46,080
Container orchestration now
comes in the form of code spaces

101
00:04:46,080 --> 00:04:49,560
online, testing via GitHub
actions, version control via Git

102
00:04:49,560 --> 00:04:51,030
itself, and more.

103
00:04:51,030 --> 00:04:53,910
And so within each of these
environments over the years

104
00:04:53,910 --> 00:04:57,090
have we also developed a suite of
command line tools and later web tools

105
00:04:57,090 --> 00:05:00,740
as well that leverage these primitives
to solve more problems for students

106
00:05:00,740 --> 00:05:04,430
in order to uplift them in
the experience of writing code

107
00:05:04,430 --> 00:05:05,210
on their own.

108
00:05:05,210 --> 00:05:08,570
Submit50, for instance, was one of the
earliest of these tools, a command line

109
00:05:08,570 --> 00:05:11,870
tool that, at the end of the day, just
let students submit their code using

110
00:05:11,870 --> 00:05:14,780
Git underneath the hood but without
having to understand or wrestle

111
00:05:14,780 --> 00:05:18,262
with any of the mechanics of Git
itself in that first week of the class,

112
00:05:18,262 --> 00:05:19,970
where they're still
just learning, hello,

113
00:05:19,970 --> 00:05:22,520
world, in C or some other language.

114
00:05:22,520 --> 00:05:26,180
But for the backend then of submit50 are
students really, at the end of the day,

115
00:05:26,180 --> 00:05:28,910
pushing to a repo in our
own GitHub organization.

116
00:05:28,910 --> 00:05:33,560
And they reference it by way of a
path, like owner/repo/branch/path that

117
00:05:33,560 --> 00:05:37,760
submit50 then uses to infer to
where to push the code for students.

118
00:05:37,760 --> 00:05:40,580
In that repo meanwhile is there
just a simple configuration

119
00:05:40,580 --> 00:05:45,200
file that flags that path as a valid
submission path that students can use.

120
00:05:45,200 --> 00:05:48,890
And within that YAML file might there
be the equivalent in YAML of something

121
00:05:48,890 --> 00:05:51,440
like a gitignore that
prescribes exactly what files

122
00:05:51,440 --> 00:05:53,390
we do and don't want them to submit.

123
00:05:53,390 --> 00:05:55,520
Now, underneath the hood
all submit50 is doing

124
00:05:55,520 --> 00:05:58,920
is essentially git add, git
commit, git push, and the like.

125
00:05:58,920 --> 00:06:00,740
And even for resubmitting
do we even avoid

126
00:06:00,740 --> 00:06:04,280
merge conflicts or the potential thereof
altogether by doing a bit of trickery

127
00:06:04,280 --> 00:06:07,490
as well so that, again, for students
early on these get completely

128
00:06:07,490 --> 00:06:08,510
abstracted away.

129
00:06:08,510 --> 00:06:12,110
But towards terms end, particularly when
they start to collaborate potentially

130
00:06:12,110 --> 00:06:15,350
on final projects, can we begin to
take away some of these training wheels

131
00:06:15,350 --> 00:06:17,880
and have them use the
native tools themselves.

132
00:06:17,880 --> 00:06:21,500
Meanwhile, to distribute students
own repositories and within it

133
00:06:21,500 --> 00:06:24,650
starter code, have we been using
GitHub Classroom most recently, whereby

134
00:06:24,650 --> 00:06:27,830
students accept an assignment
which has the result of copying

135
00:06:27,830 --> 00:06:31,850
a template repository for them into our
own org that they and perhaps their TA

136
00:06:31,850 --> 00:06:34,280
then have read and/or write access to.

137
00:06:34,280 --> 00:06:37,760
Underneath the hood meanwhile do we then
leverage web hooks or the equivalent

138
00:06:37,760 --> 00:06:39,860
thereof to use GitHub
actions, whereby we

139
00:06:39,860 --> 00:06:42,500
export a prescribed format
for GitHub Classroom

140
00:06:42,500 --> 00:06:45,650
in the form of an autograding.json
file that actually automates

141
00:06:45,650 --> 00:06:48,657
the process of running a suite of
correctness tests, which, in our case,

142
00:06:48,657 --> 00:06:50,990
happen to use our own command
line tool but could really

143
00:06:50,990 --> 00:06:53,010
be any unit testing tool or the like.

144
00:06:53,010 --> 00:06:55,010
And for more qualitative
feedback do we leverage

145
00:06:55,010 --> 00:06:57,050
the web UI to provide
line by line comments

146
00:06:57,050 --> 00:06:59,300
as might be the case with a
teaching assistant working

147
00:06:59,300 --> 00:07:01,355
more closely with their own student.

148
00:07:01,355 --> 00:07:03,230
And now in our case for
those automated tests

149
00:07:03,230 --> 00:07:07,640
do we use our own tool, check50, which
essentially just wraps a unit testing

150
00:07:07,640 --> 00:07:09,600
suite for C, and Python, and the like.

151
00:07:09,600 --> 00:07:11,900
But that too ultimately is
backed by a GitHub repo.

152
00:07:11,900 --> 00:07:16,430
So a student would run a command like
check50 owner/repo/branch/path that

153
00:07:16,430 --> 00:07:20,990
simply indicates where in the web to
go get those correctness tests from

154
00:07:20,990 --> 00:07:24,110
in order to run them locally
on the student's own code.

155
00:07:24,110 --> 00:07:27,393
Now, in the graphical context, we
also have provided students and built

156
00:07:27,393 --> 00:07:30,560
on top of these primitives something
we call see CS50 Lab, which essentially

157
00:07:30,560 --> 00:07:34,940
embeds a text editor and terminal window
alongside a rendered markdown file,

158
00:07:34,940 --> 00:07:37,640
thereby enabling us
and any teacher online

159
00:07:37,640 --> 00:07:41,105
to create in their own GitHub repository
a learning experience for students.

160
00:07:41,105 --> 00:07:43,730
That might have some starter
code, might have some correctness,

161
00:07:43,730 --> 00:07:46,910
and might also have a narrative
alongside it that ultimately lives

162
00:07:46,910 --> 00:07:52,250
in a repository indicated by, again,
owner/repo/branch/path inside of which

163
00:07:52,250 --> 00:07:54,890
too might be a YAML file for
some minimal configuration,

164
00:07:54,890 --> 00:07:57,170
the readme file itself,
and the starter code.

165
00:07:57,170 --> 00:08:00,200
And within that YAML file would
just be a bunch of key value pairs

166
00:08:00,200 --> 00:08:03,380
that tell our tool how to configure
the web environment exactly

167
00:08:03,380 --> 00:08:06,380
as that teacher wants
for their own students.

168
00:08:06,380 --> 00:08:09,350
Also, graphically as well within
these most recent web tools,

169
00:08:09,350 --> 00:08:11,510
like the IDE and now VS
Code, have we wrapped

170
00:08:11,510 --> 00:08:14,750
the process of just starting the
graphical debugger by a tool called

171
00:08:14,750 --> 00:08:19,220
debug50 so that when students compile
and then run their code through debug50

172
00:08:19,220 --> 00:08:22,910
does it trigger essentially
a launch.json file in VS Code

173
00:08:22,910 --> 00:08:26,033
to get dynamically generated and
configured so that the debugger is

174
00:08:26,033 --> 00:08:28,700
right there up and running without
students manually configuring

175
00:08:28,700 --> 00:08:29,870
anything themselves.

176
00:08:29,870 --> 00:08:33,020
If you're familiar with rubber duck
debugging, the process of talking

177
00:08:33,020 --> 00:08:36,500
through your logical problems in hopes
of hearing any illogic with a rubber

178
00:08:36,500 --> 00:08:38,870
duck or an inanimate
object, similarly, have

179
00:08:38,870 --> 00:08:41,090
we built a VS Code extension
that allows students

180
00:08:41,090 --> 00:08:43,909
to talk via chat with
a virtual rubber duck

181
00:08:43,909 --> 00:08:47,090
giving them hopefully a bit of
prompt for finding their problems.

182
00:08:47,090 --> 00:08:49,700
And then underlying all of
these tools is not only Docker

183
00:08:49,700 --> 00:08:52,580
but now a devcontainer.json
file, which allows

184
00:08:52,580 --> 00:08:54,650
us to specify a whole
suite of configurations

185
00:08:54,650 --> 00:08:58,070
for students, like the extensions to
preinstall, the Docker image to use,

186
00:08:58,070 --> 00:09:00,830
and the settings that can
customize the graphical user

187
00:09:00,830 --> 00:09:02,780
interface they're ultimately using.

188
00:09:02,780 --> 00:09:05,510
Behind the scenes, meanwhile,
do we use a suite of REST APIs

189
00:09:05,510 --> 00:09:07,350
in order to automate a
lot of the processes,

190
00:09:07,350 --> 00:09:09,380
especially with regard to authorization.

191
00:09:09,380 --> 00:09:11,690
But ultimately, these
and more tools are all

192
00:09:11,690 --> 00:09:14,660
freely available to students and
teachers alike and documented

193
00:09:14,660 --> 00:09:16,730
at cs50.readthedocs.io.

194
00:09:16,730 --> 00:09:20,510
The course's themselves are all
freely available at edx.org/cs50.

195
00:09:20,510 --> 00:09:24,070
And indeed, this was CS50.

196
00:09:24,070 --> 00:09:25,000