1
00:00:00,000 --> 00:00:00,032


2
00:00:00,032 --> 00:00:01,740
SPEAKER 1: No longer
in the weeks to come

3
00:00:01,740 --> 00:00:05,190
are we going to be focusing
on the command line alone.

4
00:00:05,190 --> 00:00:08,280
DAVID J MALAN: So this has been an
evolution, I think, in the classroom.

5
00:00:08,280 --> 00:00:10,410
Increasingly, in recent
years, have we tried

6
00:00:10,410 --> 00:00:13,380
to-- have I tried to keep pace
with certain trends in industry

7
00:00:13,380 --> 00:00:14,730
and programming paradigms.

8
00:00:14,730 --> 00:00:18,390
And as MVC became more and
more popular in Python,

9
00:00:18,390 --> 00:00:21,420
and PHP, and other languages, still,
did we weave it into the course.

10
00:00:21,420 --> 00:00:23,902
So we weren't just
using raw PHP years ago,

11
00:00:23,902 --> 00:00:25,860
and we're not just using
Python out of the box.

12
00:00:25,860 --> 00:00:28,651
We're actually standing on the
shoulders of very popular frameworks

13
00:00:28,651 --> 00:00:30,780
that actually do solve
very well-defined problems.

14
00:00:30,780 --> 00:00:34,860
And the reality is to take away
some of those abstractions and some

15
00:00:34,860 --> 00:00:39,870
of that structure, we
would be doing students,

16
00:00:39,870 --> 00:00:41,970
I think, a disservice to
not show them how they're

17
00:00:41,970 --> 00:00:43,720
going to do things in the real world.

18
00:00:43,720 --> 00:00:45,630
And so we instead use
something in Python

19
00:00:45,630 --> 00:00:48,090
like Flask, which is a very
popular micro framework--

20
00:00:48,090 --> 00:00:51,180
micro framework in that it's not
the bloated framework that some

21
00:00:51,180 --> 00:00:52,900
of the larger things are.

22
00:00:52,900 --> 00:00:56,070
But it allows us to get certain jobs
done like routing, for instance,

23
00:00:56,070 --> 00:00:59,280
and other certain functionality like
parsing the [? HTTP ?] requests that

24
00:00:59,280 --> 00:00:59,779
come in.

25
00:00:59,779 --> 00:01:00,750
And so that's nice.

26
00:01:00,750 --> 00:01:03,750
DOUG LLOYD: Were you concerned about
using a micro framework, as opposed

27
00:01:03,750 --> 00:01:05,370
to one of the more full fledged--

28
00:01:05,370 --> 00:01:08,119
that's not really the right term,
but one of the larger frameworks

29
00:01:08,119 --> 00:01:09,720
that might have some more features?

30
00:01:09,720 --> 00:01:11,320
DAVID J MALAN: No, quite the opposite.

31
00:01:11,320 --> 00:01:14,580
I mean, I either wanted to give
them nothing which felt extreme,

32
00:01:14,580 --> 00:01:19,567
or the least amount of structure
possible to get real work done.

33
00:01:19,567 --> 00:01:21,900
And so by contrast, we could
have something like Django,

34
00:01:21,900 --> 00:01:25,066
but you just get so much more with that,
and you get the whole kitchen sink.

35
00:01:25,066 --> 00:01:29,069
And I think you lose the appreciation
of what you have to build

36
00:01:29,069 --> 00:01:30,860
and what the framework
is building for you.

37
00:01:30,860 --> 00:01:32,776
I think when you're new
to the language itself

38
00:01:32,776 --> 00:01:35,310
as most students were with
Python, you might not necessarily

39
00:01:35,310 --> 00:01:37,892
grasp the distinction between
what is Python the language,

40
00:01:37,892 --> 00:01:39,600
and what is Django
the framework in terms

41
00:01:39,600 --> 00:01:41,430
of what functions are available to you.

42
00:01:41,430 --> 00:01:44,225
So Flask seemed to
strike the right balance

43
00:01:44,225 --> 00:01:47,100
DOUG LLOYD: I mean it was definitely
an overhaul of [? P SET 7. ?] We

44
00:01:47,100 --> 00:01:49,980
used to be that all of the
functionalities of P SET 7,

45
00:01:49,980 --> 00:01:54,930
were mock e-trade like website where
we have students buying, and selling,

46
00:01:54,930 --> 00:01:55,740
and quoting stocks.

47
00:01:55,740 --> 00:01:56,580
And all of those--

48
00:01:56,580 --> 00:01:59,330
all that functionality used to be
the individual files of students

49
00:01:59,330 --> 00:02:02,730
that write [? by.php ?] and implement
all of the functionality required

50
00:02:02,730 --> 00:02:08,520
to purchase a stock in its
own file, but in Python, we

51
00:02:08,520 --> 00:02:11,190
had to kind of reinvent the
way the assignment was set up.

52
00:02:11,190 --> 00:02:11,940
DAVID J MALAN: It simplified it a bit.

53
00:02:11,940 --> 00:02:13,410
I think it kept all
of the controller code

54
00:02:13,410 --> 00:02:16,326
that students were writing in one
place, which I think was compelling.

55
00:02:16,326 --> 00:02:20,940
And what's been important to me too,
is like we realize that many students--

56
00:02:20,940 --> 00:02:23,250
most students perhaps who
do web based final projects,

57
00:02:23,250 --> 00:02:26,310
are going to almost always
base their own project

58
00:02:26,310 --> 00:02:29,952
on [? P SET 7's ?] distribution code
or [? P SET 8, ?] and I really did want

59
00:02:29,952 --> 00:02:32,910
the code were giving students and
the blanks they were filling in to be

60
00:02:32,910 --> 00:02:36,640
represented of how they should start
their own projects once they graduate

61
00:02:36,640 --> 00:02:37,620
from CS50.

62
00:02:37,620 --> 00:02:41,850
And so I mean we really obsessed over
every line of code and the distribution

63
00:02:41,850 --> 00:02:43,950
code, making sure that we
can justify everything.

64
00:02:43,950 --> 00:02:45,991
And making sure that it
really was representative

65
00:02:45,991 --> 00:02:48,720
of how you might start a
general web application that

66
00:02:48,720 --> 00:02:50,490
has both the back end and a front end.

67
00:02:50,490 --> 00:02:52,948
DOUG LLOYD: But what's also
interesting about [? P SET 7 ?]

68
00:02:52,948 --> 00:02:55,622
is that because we had a
problem set in Python before,

69
00:02:55,622 --> 00:02:58,080
we were actually able to take
off some more of the training

70
00:02:58,080 --> 00:02:59,871
than we would have
otherwise been able to--

71
00:02:59,871 --> 00:03:04,279
We always had to leave on in the
PHP iteration of this problem set.

72
00:03:04,279 --> 00:03:05,820
DAVID J MALAN: That was nice I think.

73
00:03:05,820 --> 00:03:07,611
I think the price we
pay is something to co

74
00:03:07,611 --> 00:03:11,100
mingle our discussion and
lecture of Python itself

75
00:03:11,100 --> 00:03:13,760
with Python at the command line
in Python in a web context.

76
00:03:13,760 --> 00:03:14,468
DOUG LLOYD: True.

77
00:03:14,468 --> 00:03:17,400
DAVID J MALAN: And so this
web programming in Python

78
00:03:17,400 --> 00:03:18,870
could potentially stand on its own.

79
00:03:18,870 --> 00:03:20,370
We didn't have that additional week.

80
00:03:20,370 --> 00:03:23,430
And so we really focused the last
portion of this particular week

81
00:03:23,430 --> 00:03:26,440
on the web application of the
same, but I think that's OK.

82
00:03:26,440 --> 00:03:28,320
Because at the end of
the day, we already

83
00:03:28,320 --> 00:03:30,600
have from the previous week what--

84
00:03:30,600 --> 00:03:33,600
from previous weeks like what the
internet is, what's HTTP and TCP.

85
00:03:33,600 --> 00:03:35,309
So we're already layering and layering.

86
00:03:35,309 --> 00:03:38,100
So to start with Python, and then
end with a bit of web programming

87
00:03:38,100 --> 00:03:40,230
in Python, I feel like it's
the right natural progression.

88
00:03:40,230 --> 00:03:42,090
And then you go one step
further of course a week later,

89
00:03:42,090 --> 00:03:43,506
with JavaScript and the front end.

90
00:03:43,506 --> 00:03:47,240
And so it kind of works,
though we could certainly--

91
00:03:47,240 --> 00:03:51,000
it would certainly be nice to have
even more time for every detail.

92
00:03:51,000 --> 00:03:54,485
DOUG LLOYD: If only
we had 30 more weeks.

93
00:03:54,485 --> 00:03:54,985