1 00:00:05,931 --> 00:00:07,341 [MUSIC]. We've come to the last section of the 2 00:00:07,341 --> 00:00:10,428 course. this at this stage we 're going to talk 3 00:00:10,428 --> 00:00:17,080 about how higher level languages like Java are implemented in computer systems. 4 00:00:17,080 --> 00:00:21,230 And we're going to leverage the knowledge we have about how things are done using 5 00:00:21,230 --> 00:00:24,800 C. what we focused on as our language so 6 00:00:24,800 --> 00:00:28,620 far. So we're going to start that process by 7 00:00:28,620 --> 00:00:32,560 going back to the very beginning and sort of replaying the course at a much faster 8 00:00:32,560 --> 00:00:36,525 pace. And not just talk about how data 9 00:00:36,525 --> 00:00:42,700 representations, pointers, and references procedures or methods in the case of Java 10 00:00:42,700 --> 00:00:46,180 are represented in a, in this higher level language. 11 00:00:46,180 --> 00:00:50,840 And we're also going to talk a little bit about virtual machines, a special runtime 12 00:00:50,840 --> 00:00:55,700 system we use for languages like java which are interpretted and not compiled 13 00:00:55,700 --> 00:00:59,020 directly. So we'll get, as I said we're going to do 14 00:00:59,020 --> 00:01:03,740 this really quick this time. because we have a lot more background to 15 00:01:03,740 --> 00:01:08,950 base things on now. So let's start turning out attention to 16 00:01:08,950 --> 00:01:11,540 Java. Before we do that just to make a meta 17 00:01:11,540 --> 00:01:16,415 point about this this lecture, is none of things we're going to talk about here are 18 00:01:16,415 --> 00:01:20,240 guaranteed by Java. What we're doing here is just trying to 19 00:01:20,240 --> 00:01:25,990 understand one implementation of a higher level language so we can get a sense for 20 00:01:25,990 --> 00:01:33,020 what the basic underlying concepts are. useful ways to think about our programs 21 00:01:33,020 --> 00:01:38,970 and what is going on underneath, but none of, none of what you hear in this set of 22 00:01:38,970 --> 00:01:42,420 videos. In this section should you take to be the 23 00:01:42,420 --> 00:01:45,810 gospel truth about a particular implementation. 24 00:01:47,489 --> 00:01:51,370 Let's begin by looking at how data is represented in Java. 25 00:01:51,370 --> 00:01:56,220 integers, floats doubles, all those things all those numbers that we were 26 00:01:56,220 --> 00:02:00,480 looking at before in the beginning of the course and see or represented exactly the 27 00:02:00,480 --> 00:02:03,680 same way in Java. there's no differences there. 28 00:02:03,680 --> 00:02:08,520 And yes a Java has pointers as well, however, there are some differences 29 00:02:08,520 --> 00:02:11,380 there. in Java pointers are referred to as 30 00:02:11,380 --> 00:02:15,230 references. And there not quite as general as a, C 31 00:02:15,230 --> 00:02:18,806 pointers are. remember with C we could point to any 32 00:02:18,806 --> 00:02:23,990 byte in memory, any arbitrary address. We'll see that in Java our references are 33 00:02:23,990 --> 00:02:28,140 always pointing to the start of an object, the start of a data structure. 34 00:02:28,140 --> 00:02:36,010 So not quite as general as C pointers. Just like in C, null is typically 35 00:02:36,010 --> 00:02:39,600 represented as a 0. and we're going to review characters and 36 00:02:39,600 --> 00:02:45,570 strings, arrays and objects. So let's start with looking at characters 37 00:02:45,570 --> 00:02:51,550 and strings in Java. the first thing to to note is that the 38 00:02:51,550 --> 00:02:57,750 character set that a Java uses has two bites per character rather then one. 39 00:02:57,750 --> 00:03:01,904 Rather then using the Latin alphabet focused ASCII encoding. 40 00:03:01,904 --> 00:03:07,300 we use two byte unit code characters which allow us to represent many more the 41 00:03:07,300 --> 00:03:12,640 world's alphabets. So it's much more general modern computer 42 00:03:12,640 --> 00:03:17,030 code for all the character sets that are out there. 43 00:03:17,030 --> 00:03:24,510 So the other difference is that unlike in C where strings are bounded by a special 44 00:03:24,510 --> 00:03:29,770 byte that has a null or a zero in it as in this example here. 45 00:03:29,770 --> 00:03:33,150 You'll notice the strings ends when we hit the zero. 46 00:03:33,150 --> 00:03:36,980 We don't know in advance how long the string is going to be. 47 00:03:36,980 --> 00:03:41,960 We often have to read the entire string. And keep checking every byte as we read 48 00:03:41,960 --> 00:03:45,120 it to know when we get to the end. when we get to that zero. 49 00:03:45,120 --> 00:03:49,590 On the other hand, in Java, the first four bytes are an integer that has the 50 00:03:49,590 --> 00:03:53,220 length of the string represented as as a number. 51 00:03:53,220 --> 00:03:59,383 in this case the number six for the six pairs of Unicode bytes. 52 00:03:59,383 --> 00:04:08,620 for the six characters in CSE351. Okay, so, a, with a, with java at the 53 00:04:08,620 --> 00:04:11,450 very beginning we know the length of the string already. 54 00:04:11,450 --> 00:04:16,040 That first thing we read is an integer that tells us how long it is and how far 55 00:04:16,040 --> 00:04:20,580 we'll need to go. The same approach carries through into 56 00:04:20,580 --> 00:04:25,330 arrays. you'll notice that arrays in Java have an 57 00:04:25,330 --> 00:04:29,140 integer at the very beginning that tells us how many elements there are in the 58 00:04:29,140 --> 00:04:32,929 array. Another difference in arrays in Java, is 59 00:04:32,929 --> 00:04:39,480 that they're already initialized to 0. and unlike in C, where the array is 60 00:04:39,480 --> 00:04:43,620 placed at some locations in memory and it's starting values are whatever was 61 00:04:43,620 --> 00:04:46,490 there. Whatever bits were currently at those 62 00:04:46,490 --> 00:04:50,980 memory locations. instead in Java there's a special effort 63 00:04:50,980 --> 00:04:54,940 made to put zeroes there to start. So that we know that whenever we read a 64 00:04:54,940 --> 00:04:57,638 value for the first time, it will in fact, be a zero. 65 00:04:57,638 --> 00:05:06,550 That nice first value there, the length of the array is accessed by a special 66 00:05:06,550 --> 00:05:12,110 method called length. That we can call on the array and that 67 00:05:12,110 --> 00:05:17,060 will return the value of that field. So that anytime we can find out how big 68 00:05:17,060 --> 00:05:22,200 this array is. And why is that there, what can that do 69 00:05:22,200 --> 00:05:25,730 for us. Well what's, what happens in Java is 70 00:05:25,730 --> 00:05:29,235 every time we access an array, every time we index an array. 71 00:05:29,235 --> 00:05:35,300 we actually go and take, you know in this case, for example, let's say we had 72 00:05:35,300 --> 00:05:40,610 written array three. we will actually go and take that value 73 00:05:40,610 --> 00:05:47,838 three and compare it against this one stored at the very first position in the 74 00:05:47,838 --> 00:05:50,960 object. And make sure that we're not addressing a 75 00:05:50,960 --> 00:05:56,900 location that's out of bounds. an, an element of the array that is 76 00:05:56,900 --> 00:06:02,460 actually defined. in this case 3 is definitely okay, it 77 00:06:02,460 --> 00:06:07,750 really corresponds to this element right here and we can go ahead and read that. 78 00:06:07,750 --> 00:06:12,778 If we had put a 7 here then that would've been somewhere out here. 79 00:06:12,778 --> 00:06:18,170 in C and we would've just generated an, an address and read whatever was at that 80 00:06:18,170 --> 00:06:21,510 memory location. In Java on the other hand, we compare 81 00:06:21,510 --> 00:06:25,110 that seven against the five and say, well we're clearly out of bounds. 82 00:06:25,110 --> 00:06:30,550 Let's throw an exception and and cause the program to stop because we're clearly 83 00:06:30,550 --> 00:06:37,520 addressing something in an incorrect way. So this is one nice advantage of writing 84 00:06:37,520 --> 00:06:41,130 Java programs that we get these extra bound checks for free. 85 00:06:41,130 --> 00:06:45,800 That means that there's extra code executing that we didn't bother to write 86 00:06:45,800 --> 00:06:51,460 to check that bound but was inserted for us automatically in the Java 87 00:06:51,460 --> 00:06:54,464 implementation to do that bounds checking. 88 00:06:54,464 --> 00:07:03,010 Okay, let's take a look at data structures in Java, or objects as they're 89 00:07:03,010 --> 00:07:10,240 referred to. the difference between structs in C and 90 00:07:10,240 --> 00:07:16,490 objects in Java, is that objects can only include primitive data types. 91 00:07:16,490 --> 00:07:21,310 Not composite data types. So for example, you'll remember maybe 92 00:07:21,310 --> 00:07:25,790 this example that we used before. Here we have a struct that has an integer 93 00:07:25,790 --> 00:07:31,350 as its first element, and then an array of three ints as its second element. 94 00:07:31,350 --> 00:07:37,550 That's not a primitive structure or primitive data type, that's a composite 95 00:07:37,550 --> 00:07:40,930 data type, it actually consists of three integers. 96 00:07:40,930 --> 00:07:46,430 And then another primitive data type that is a pointer, okay? 97 00:07:46,430 --> 00:07:54,735 when we access these in C, what we can do is allocate some space for for the, the 98 00:07:54,735 --> 00:08:00,360 struct. Get a pointer to that region of memory, 99 00:08:00,360 --> 00:08:06,130 and then we can refer to it using that arrow notation which says, hey take that 100 00:08:06,130 --> 00:08:11,270 pointer, dereference it, offset by the amount that you need to offset to get to 101 00:08:11,270 --> 00:08:17,500 the various pieces of the, of the struct and assign a new value to them. 102 00:08:17,500 --> 00:08:22,390 And you'll notice that we can not only reference a, but we can reference an 103 00:08:22,390 --> 00:08:27,590 element within a, and assign a value to one element of the array. 104 00:08:27,590 --> 00:08:32,425 here, we're assigning an address of another struct to that pointer. 105 00:08:32,425 --> 00:08:36,270 Okay. So, in memory, this looks like. 106 00:08:37,960 --> 00:08:42,620 An integer, followed by three more integers that represent the array, a, and 107 00:08:42,620 --> 00:08:47,840 then another four bytes that are the pointer, that is the third element. 108 00:08:47,840 --> 00:08:52,640 Let's see what that looks like in Java. You'll notice that in Java, yes, we do 109 00:08:52,640 --> 00:09:01,150 have the integer as the first element and the reference to another object a pointer 110 00:09:01,150 --> 00:09:06,630 like thing, as the third element, but in the middle we don't have the actual array 111 00:09:06,630 --> 00:09:13,490 of three integers, we have a pointer to a new structure that is an array of three 112 00:09:13,490 --> 00:09:18,394 integers, so rather than actually representing the three integers there. 113 00:09:18,394 --> 00:09:22,540 We actually just represent a pointer to the array. 114 00:09:23,870 --> 00:09:27,180 So to, let's just take a look at that map. 115 00:09:27,180 --> 00:09:32,400 So here in this case, our memory has our four byte integer at the beginning, our 116 00:09:32,400 --> 00:09:40,670 four byte pointer to an array and another four byte reference to another object. 117 00:09:40,670 --> 00:09:46,010 So that pointer a is not the actual array, but rather points off to a region 118 00:09:46,010 --> 00:09:50,850 of memory where the array is stored. Here are the three ints, but remember 119 00:09:50,850 --> 00:09:53,200 they're preceded by the length of the array. 120 00:09:53,200 --> 00:09:56,310 So there's an extra integer there at the beginning. 121 00:09:56,310 --> 00:09:59,030 That tells us how many elements there are to this array. 122 00:09:59,030 --> 00:10:06,030 When we reference them in code, when we reference the struct in code you'll 123 00:10:06,030 --> 00:10:08,130 notice theres a slightly different syntax. 124 00:10:08,130 --> 00:10:15,520 We don't bother doing malloc to call the memory allocation to get us a space in 125 00:10:15,520 --> 00:10:20,270 memory. We use the keyword new which is actually 126 00:10:20,270 --> 00:10:26,080 something that does that malloc for us and creates a new space in memory, in 127 00:10:26,080 --> 00:10:33,720 this case two new spaces in memory. One for the first level of the struct and 128 00:10:33,720 --> 00:10:39,060 then one for one of, another area for the composite data type that makes up that 129 00:10:39,060 --> 00:10:44,110 that struct. That extra array that has to be created 130 00:10:44,110 --> 00:10:47,499 and it initializes these, this pointer to point to that array. 131 00:10:49,220 --> 00:10:53,520 The references to the elements are very similar to what we did in C, but you'll 132 00:10:53,520 --> 00:10:55,420 notice that we've changed the notation a little bit. 133 00:10:55,420 --> 00:10:59,910 Instead of the arrow, we just used a dot. It's it's a little bit shorter, easier to 134 00:10:59,910 --> 00:11:04,710 read, and since we're always using a reference that's to the start of an 135 00:11:04,710 --> 00:11:08,840 object. we implicitly do the do reference so we 136 00:11:08,840 --> 00:11:16,020 don't bother with that extra step, and can just use the dot notation. 137 00:11:16,020 --> 00:11:20,250 Alright, so you've seen a, a little bit about these pointers, let's see, a, what 138 00:11:20,250 --> 00:11:24,490 that implications that has. So a pointer in C, again, can point to 139 00:11:24,490 --> 00:11:28,150 any memory address. In, in Java though, it can only point to 140 00:11:28,150 --> 00:11:31,980 the beginning of an object. only it's first element, not, it cannot 141 00:11:31,980 --> 00:11:35,060 point to the middle of it. So what implications does this have? 142 00:11:35,060 --> 00:11:39,270 Well you, here's our struct again and you'll notice here we're calling a 143 00:11:39,270 --> 00:11:44,150 function with one element of the array that makes up that struct and we can do 144 00:11:44,150 --> 00:11:49,560 that with a single expression. We go to that struct, look at its a 145 00:11:49,560 --> 00:11:53,040 array, and then the second element of that array. 146 00:11:53,040 --> 00:11:58,040 And the address of that element, and can pass that on to the function. 147 00:11:58,040 --> 00:12:03,380 So in this case what we're doing is passing a pointer to that function, some 148 00:12:03,380 --> 00:12:09,860 fun As an argument and that ar, that argument points to a particular integer 149 00:12:09,860 --> 00:12:13,310 in the middle of that array. In Java, we can't do that. 150 00:12:13,310 --> 00:12:16,060 Remember our pointers? Our references can only go to the 151 00:12:16,060 --> 00:12:20,430 beginning of an object so, we actually have to specify it in two parts. 152 00:12:20,430 --> 00:12:26,158 First, we have to say the array Address and then, which element of the array. 153 00:12:26,158 --> 00:12:30,210 we, because we can't point to the middle of the array directly. 154 00:12:30,210 --> 00:12:35,430 So remember, that's what our struct looks like in our object looks like in memory. 155 00:12:35,430 --> 00:12:41,240 These two discontiguous parts, and we cannot generate a pointer to the middle 156 00:12:41,240 --> 00:12:47,400 of that array so the way we have to do it is by pointing first to the array by 157 00:12:47,400 --> 00:12:53,690 getting that pointer stored at this location, and then providing an index 158 00:12:53,690 --> 00:12:58,845 that tells us which element to the array we want, and that index will be compared 159 00:12:58,845 --> 00:13:03,650 against the length of the array and checked for us. 160 00:13:03,650 --> 00:13:06,630 to make sure that we're not over stepping the bounds.