1
00:00:06,692 --> 00:00:09,470
NATE HARDISON: In programming, a library is a collection of

2
00:00:09,470 --> 00:00:11,380
related prewritten code.

3
00:00:11,380 --> 00:00:14,350
Libraries are how we, as programmers, share common and

4
00:00:14,350 --> 00:00:16,930
helpful code with each other, with the different programs we

5
00:00:16,930 --> 00:00:19,470
write, and even with the different processes running at

6
00:00:19,470 --> 00:00:21,380
the same time on our computers.

7
00:00:21,380 --> 00:00:23,000
>> Let's explore a bit.

8
00:00:23,000 --> 00:00:25,070
One of the more common functions that you've probably

9
00:00:25,070 --> 00:00:27,500
been using is the printf function.

10
00:00:27,500 --> 00:00:29,820
Now, printf isn't some magic function that's been

11
00:00:29,820 --> 00:00:31,670
hard-coded into the computer--

12
00:00:31,670 --> 00:00:34,320
rather, it's part of the C Standard Libraries, which are

13
00:00:34,320 --> 00:00:36,080
a collection of functions that comes with the

14
00:00:36,080 --> 00:00:38,050
C programming language.

15
00:00:38,050 --> 00:00:40,920
Since printf isn't built into the computer, that means that

16
00:00:40,920 --> 00:00:43,520
there was some programmer who actually went in and wrote the

17
00:00:43,520 --> 00:00:46,310
printf function and bundled it with the rest of the standard

18
00:00:46,310 --> 00:00:48,880
libraries so that future programmers wouldn't have to

19
00:00:48,880 --> 00:00:50,310
duplicate the effort.

20
00:00:50,310 --> 00:00:52,620
And we're sure thankful that that's the case, because check

21
00:00:52,620 --> 00:00:54,710
out how much code it actually takes to implement printf.

22
00:00:59,590 --> 00:01:02,860
>> The C Standard Libraries, of which printf is a part, are

23
00:01:02,860 --> 00:01:04,480
one of the most important tools you

24
00:01:04,480 --> 00:01:06,770
have in tackling CS50.

25
00:01:06,770 --> 00:01:08,780
In addition to the input-output library where

26
00:01:08,780 --> 00:01:11,370
printf lives, there are a bunch of other libraries that

27
00:01:11,370 --> 00:01:14,230
you'll probably use over the course of CS50.

28
00:01:14,230 --> 00:01:16,730
For example, the string library has functions to

29
00:01:16,730 --> 00:01:20,960
compare C strings, get their length, and concatenate them.

30
00:01:20,960 --> 00:01:22,410
And the math library gives you helpful

31
00:01:22,410 --> 00:01:25,062
constants, like pi and e.

32
00:01:25,062 --> 00:01:27,650
Now, compared to the standard libraries of other programming

33
00:01:27,650 --> 00:01:31,190
languages, the C Standard Libraries are really small,

34
00:01:31,190 --> 00:01:34,060
but they still have a significant amount of stuff.

35
00:01:34,060 --> 00:01:36,370
And that's not to say that the Standard Libraries are the

36
00:01:36,370 --> 00:01:39,020
only C libraries; there are many more out there in the

37
00:01:39,020 --> 00:01:42,460
world for you to use, including the CS50 Library,

38
00:01:42,460 --> 00:01:45,150
cartography libraries with encryption and decryption

39
00:01:45,150 --> 00:01:48,000
functions, libraries with functions to encode and

40
00:01:48,000 --> 00:01:50,970
play back video, and so on.

41
00:01:50,970 --> 00:01:53,190
>> This brings us to an important aside.

42
00:01:53,190 --> 00:01:55,500
One of the essential skills of engineering is knowing the

43
00:01:55,500 --> 00:01:58,090
tools available in your toolbox so that you can use

44
00:01:58,090 --> 00:01:59,850
them and avoid reinventing the wheel unless

45
00:01:59,850 --> 00:02:01,510
you have to do so.

46
00:02:01,510 --> 00:02:04,120
As programmers, libraries are one of the most important

47
00:02:04,120 --> 00:02:05,520
tools we have.

48
00:02:05,520 --> 00:02:08,520
When you find yourself facing a task that seems tedious or

49
00:02:08,520 --> 00:02:11,940
commonplace, like finding the length of a C string, you

50
00:02:11,940 --> 00:02:14,020
should ask yourself if it might be the case that

51
00:02:14,020 --> 00:02:16,520
someone's already done the same thing before.

52
00:02:16,520 --> 00:02:17,760
Chances are that you could find some

53
00:02:17,760 --> 00:02:20,110
help within a library.

54
00:02:20,110 --> 00:02:23,260
>> So in technical terms, a library is a binary file that

55
00:02:23,260 --> 00:02:26,210
has been produced by linking together a collection of

56
00:02:26,210 --> 00:02:28,880
object files using a tool called, you

57
00:02:28,880 --> 00:02:31,010
guessed it, the linker.

58
00:02:31,010 --> 00:02:33,830
Object files are those .o files that you get when you

59
00:02:33,830 --> 00:02:35,860
compile C source code.

60
00:02:35,860 --> 00:02:38,420
When programmers write libraries, they typically

61
00:02:38,420 --> 00:02:41,560
separate their C code into two types of files:

62
00:02:41,560 --> 00:02:44,880
header files and implementation files.

63
00:02:44,880 --> 00:02:49,820
A header file by convention is given the .h file extension,

64
00:02:49,820 --> 00:02:51,940
and it contains code that declares the library's

65
00:02:51,940 --> 00:02:55,120
resources that are available for you to use.

66
00:02:55,120 --> 00:02:58,830
These resources include functions, variables, structs,

67
00:02:58,830 --> 00:03:01,980
type definitions, and so on, and together, they are

68
00:03:01,980 --> 00:03:04,620
commonly referred to as the interface.

69
00:03:04,620 --> 00:03:07,280
In fact, if you're ever wondering what's in a library,

70
00:03:07,280 --> 00:03:09,580
the header files are the place to look.

71
00:03:09,580 --> 00:03:14,030
>> For example, you can pull up usr/include/stdio.h and check

72
00:03:14,030 --> 00:03:16,210
out everything you ever wanted to know about

73
00:03:16,210 --> 00:03:18,570
the standard io library.

74
00:03:18,570 --> 00:03:20,610
There's a lot of stuff in here, and it can take a bit of

75
00:03:20,610 --> 00:03:22,420
time to wade through.

76
00:03:22,420 --> 00:03:25,840
However, along with Google and the manpages, this is the

77
00:03:25,840 --> 00:03:28,060
place to go if you're wondering how the standard I/O

78
00:03:28,060 --> 00:03:30,460
library functions work.

79
00:03:30,460 --> 00:03:33,910
So one key thing to note here, though, is that header files

80
00:03:33,910 --> 00:03:36,740
typically do not include function definitions--

81
00:03:36,740 --> 00:03:40,020
that is, the actual function implementations.

82
00:03:40,020 --> 00:03:42,470
This is an illustration of an important computer science

83
00:03:42,470 --> 00:03:45,700
principle called information hiding, also referred to at

84
00:03:45,700 --> 00:03:48,960
times as encapsulation.

85
00:03:48,960 --> 00:03:52,630
>> As the user or the client of a library, you don't need to

86
00:03:52,630 --> 00:03:54,230
know about the internals of the library in

87
00:03:54,230 --> 00:03:55,380
order to use it.

88
00:03:55,380 --> 00:03:57,800
Remember seeing printf just a bit ago?

89
00:03:57,800 --> 00:04:00,275
Had you ever seen the actual code before?

90
00:04:00,275 --> 00:04:02,535
Well, I assume that you haven't, even though you might

91
00:04:02,535 --> 00:04:05,090
have used printf tons of times.

92
00:04:05,090 --> 00:04:08,110
All you had to know was the function declaration present

93
00:04:08,110 --> 00:04:10,940
in the stdio.h header file.

94
00:04:10,940 --> 00:04:14,010
Anyway, the benefit of good information hiding is that you

95
00:04:14,010 --> 00:04:16,820
and the programs you write are insulated from any changes in

96
00:04:16,820 --> 00:04:20,209
the code of the implementation of library functions.

97
00:04:20,209 --> 00:04:22,880
If the implementation of printf changes, you don't have

98
00:04:22,880 --> 00:04:26,310
to go and change all of your programs that use printf.

99
00:04:26,310 --> 00:04:28,450
>> So where does implementation go?

100
00:04:28,450 --> 00:04:31,020
Well, again, by convention, the implementation goes in a

101
00:04:31,020 --> 00:04:34,310
file with the .c file extension.

102
00:04:34,310 --> 00:04:36,820
Usually, header files and implementation files go hand

103
00:04:36,820 --> 00:04:39,380
in hand, so if you've declared a bunch of new string

104
00:04:39,380 --> 00:04:42,950
functions in mystring.h, then you'll define all of

105
00:04:42,950 --> 00:04:46,960
those functions in mystring .c.

106
00:04:46,960 --> 00:04:49,230
So then the header and implementation files are then

107
00:04:49,230 --> 00:04:52,230
compiled, creating object files that are then linked

108
00:04:52,230 --> 00:04:55,360
together to produce the binary library file.

109
00:04:55,360 --> 00:04:58,520
The library writer will then ship the binary file along

110
00:04:58,520 --> 00:05:00,720
with the header files to programmers who want to use

111
00:05:00,720 --> 00:05:02,280
the library.

112
00:05:02,280 --> 00:05:04,640
At this point, the .c implementation files aren't

113
00:05:04,640 --> 00:05:07,840
needed, except to recompile the library from scratch, so

114
00:05:07,840 --> 00:05:10,270
they're often left out.

115
00:05:10,270 --> 00:05:13,670
>> To use the C library, you have to do two things.

116
00:05:13,670 --> 00:05:16,890
First, you must #include the library's header files and the

117
00:05:16,890 --> 00:05:18,700
source code file where you wish to use

118
00:05:18,700 --> 00:05:20,530
the library's code.

119
00:05:20,530 --> 00:05:23,200
This informs the compiler of all of the declarations of

120
00:05:23,200 --> 00:05:26,890
functions, variables, structs, et cetera, present in the

121
00:05:26,890 --> 00:05:30,510
header files so that the compiler can one, insure that

122
00:05:30,510 --> 00:05:33,280
you're using the library's resources in the proper way,

123
00:05:33,280 --> 00:05:37,120
and two, generate the appropriate assembly code.

124
00:05:37,120 --> 00:05:38,910
Depending on where the library's header files are

125
00:05:38,910 --> 00:05:42,720
located, the #include syntax varies.

126
00:05:42,720 --> 00:05:44,880
If the header files are located in the system's

127
00:05:44,880 --> 00:05:46,690
default include directory--

128
00:05:46,690 --> 00:05:48,780
typically, usr/include--

129
00:05:48,780 --> 00:05:52,200
then the angle brackets syntax is used, as in #include

130
00:05:52,200 --> 00:05:53,450
.

131
00:05:57,440 --> 00:05:59,540
However, if the header files are located in the same

132
00:05:59,540 --> 00:06:02,960
directory as your own code, the double quote syntax is

133
00:06:02,960 --> 00:06:04,870
used, as in #include "mylib.h".

134
00:06:08,630 --> 00:06:11,220
>> The second thing you have to do is to link in the binary

135
00:06:11,220 --> 00:06:13,760
library file when you compile your code.

136
00:06:13,760 --> 00:06:15,810
This is a super important step.

137
00:06:15,810 --> 00:06:17,540
Remember our earlier discussion about how header

138
00:06:17,540 --> 00:06:20,200
files don't contain implementation code?

139
00:06:20,200 --> 00:06:22,200
If you forget this step, you'll get errors about

140
00:06:22,200 --> 00:06:24,990
undefined symbols referenced in your code.

141
00:06:24,990 --> 00:06:29,580
To solve this, use the -l flag followed immediately

142
00:06:29,580 --> 00:06:32,320
without a space by the name of the library.

143
00:06:32,320 --> 00:06:35,410
On many systems, including the appliance, the C Standard

144
00:06:35,410 --> 00:06:38,350
Libraries are automatically linked for you.

145
00:06:38,350 --> 00:06:41,700
However, the math library is sometimes bundled separately

146
00:06:41,700 --> 00:06:44,600
and might need to be linked with -lm.

147
00:06:44,600 --> 00:06:47,760
On the appliance, we take care of this for you, and we also

148
00:06:47,760 --> 00:06:52,170
link the CS50 library with -lcs50.

149
00:06:52,170 --> 00:06:54,540
You can see all of this on the appliance when you run make.

150
00:06:57,620 --> 00:06:59,130
>> Now you should have the knowledge you need to use

151
00:06:59,130 --> 00:07:02,150
libraries in C. Have fun exploring.

152
00:07:02,150 --> 00:07:05,880
My name is Nate Hardison, and this is CS50.