1. snick  snack 1 CPSC 121: Models of Computation 2009 Winter Term 1 Number Representation Steve Wolfman, based on notes by Patrice Belleville and others

2. Outline Prereqs, Learning Goals, and Quiz Notes Prelude: “Additive Inverse” Problems and Discussion –Clock Arithmetic and Two’s Complement –1/3 Scottish and Fractions in Binary –Programs and Numbers –Programs as Numbers Next Lecture Notes 2

3. Lecture Prerequisites Read Section 1.5 and the supplemental handout: http://www.ugrad.cs.ubc.ca/~cs121/2009W1/ Handouts/signed-binary-decimal-conversions.html http://www.ugrad.cs.ubc.ca/~cs121/2009W1/ Handouts/signed-binary-decimal-conversions.html Solve problems like Exercise Set 1.5, #1-16, 27-36, and 41-46. NOTE: the binary numbers in problems 27- 30 are signed (that is, they use the two’s complement representation scheme). Complete the open-book, untimed quiz on Vista that’s due before the next class. 3

4. Learning Goals: Pre-Class By the start of class, you should be able to: –Convert positive numbers from decimal to binary and back. –Convert positive numbers from hexadecimal to binary and back. –Take the two’s complement of a binary number. –Convert signed (either positive or negative) numbers to binary and back. –Add binary numbers. 4

5. Learning Goals: In-Class By the end of this unit, you should be able to: –Critique the choice of a digital representation scheme—including describing its strengths, weaknesses, and flaws (such as imprecise representation or overflow) —for a given type of data and purpose, such as (1) fixed-width binary numbers using a two’s complement scheme for signed integer arithmetic in computers or (2) hexadecimal for human inspection of raw binary data. 5

6. Where We Are in The Big Stories Theory How do we model computational systems? Now: showing that our logical models can connect smoothly to models of number systems. Hardware How do we build devices to compute? Now: enabling our hardware to work with data that’s more meaningful to humans. (And once we have numbers, we can represent pictures, words, sounds, and everything else!) 6

7. Outline Prereqs, Learning Goals, and Quiz Notes Prelude: “Additive Inverse” Problems and Discussion –Clock Arithmetic and Two’s Complement –1/3 Scottish and Fractions in Binary –Programs and Numbers –Programs as Numbers Next Lecture Notes 7

8. Prelude: Unsigned Integers We can choose any arrangement of Ts and Fs to represent numbers... But, we might as well choose something convenient. If we let F correspond to 0 and T to 1, then our representation is... 8 #pqr 0FFF 1FFT 2FTF 3FTT 4TFF 5TFT 6TTF 7TTT

9. Prelude: Unsigned Integers...base 2 numbers. When we represent negative numbers, the choice is also arbitrary, but may as well be convenient: Just one representation for zero Easy to tell negative from positive (or non- negative?) Basic operations easy. 9 #pqr 0000 1001 2010 3011 4100 5101 6110 7111 But... What does it mean for basic operations to be “easy”?

10. Prelude: Additive Inverse The “additive inverse” of a number x is another number y such that x + y = 0. What is the additive inverse of 3? What is the additive inverse of -7? 10 We want to be able to add signed binary numbers. We need x + -x to be 0. And, we want addition to be easy to implement.

11. Outline Prereqs, Learning Goals, and Quiz Notes Prelude: “Additive Inverse” Problems and Discussion –Clock Arithmetic and Two’s Complement –1/3 Scottish and Fractions in Binary –Programs and Numbers –Programs as Numbers Next Lecture Notes 11

12. Problem: Clock Arithmetic Problem: It’s 0500h. How many hours until midnight? Give an algorithm that requires a 24-hour clock, a level, and no arithmetic. 12 A level is a carpentry tool, essentially a straightedge that indicates when it is either horizontal or vertical.

13. Clock Arithmetic 0500 is five hours from midnight. 1900 is five hours to midnight. 5 and 19 are “additive inverses” in clock arithmetic: 5 + 19 = 0. So are any other numbers that are “across the clock” from each other. 13 That’s even true for 12. Its additive inverse is itself!

14. Clock Arithmetic Problem It’s 18 hundred. Without using numbers larger than 24 in your calculations, what time will it be 22*7 hours from now? (Don’t multiply 22 by 7!) a.0 hundred (midnight) b.4 hundred c.8 hundred d.14 hundred e.None of these 14 (Clock arithmetic is also known as modular arithmetic in mathematics.)

15. Clock Arithmetic: Food for Thought 0 -3 -6 -9 -12 If we wanted negative numbers on the clock, we’d probably put them “across the clock” from the positives. After all, if 3 + 21 is already 0, why not put -3 where 21 usually goes? 15

16. Unsigned Binary Clock Here’s a 3-bit unsigned binary clock, numbered from 0 (000) to 7 (111). 16 000 001 010 011 100 101 110 111

17. Crossing the Clock To “cross the clock”, go as many ticks left from the top as you previously went right from the top. Here’s a clock labelled with 0 (000) to 3 (011) and -1 (111) to -4 (100). 17 000 001 010 011 100 101 110 111

18. Reminder: Two’s Complement Taking two’s complement of B = b 1 b 2 b 3...b n : 1 1 1...1 - b 1 b 2 b 3...b n ---------- x 1 x 2 x 3...x n + 1 ---------- -B 18 Flip the bits Add one

19. A Different View of Two’s Complement Taking two’s complement of B = b 1 b 2 b 3...b n : 1 1 1...1 - b 1 b 2 b 3...b n ---------- x 1 x 2 x 3...x n + 1 ---------- -B 19 Flip the bits Add one 1 1 1...1 + 1 ---------- 1 0 0 0...0 - b 1 b 2 b 3...b n ---------- -B Or... Just subtract from 100...0

20. Two’s Complement vs. Crossing the Clock Two’s complement with k bits: Equivalent to subtracting from 100...000 with k 0s. “Crossing the clock” with k bits: Equivalent to subtracting from 100...000 with k 0s. 20 Two’s complement turns numbers into their “normal”, “cross-the-clock” additive inverses. 1 1 1...1 + 1 ---------- 1 0 0 0...0 - b 1 b 2 b 3...b n ---------- -B 000 001 010 011 100 101 110 111

21. Problem: Why Two’s complement? Why make the negation of 010 be 110? a.010 + 110 already equals 0. b.010 + 110 equals 1000. c.110 is the easiest negation to calculate. d.110 isn’t being used for any other purpose. e.If you invert the bits in 110 and add 1, you get 010. 000 001 010 011 100 101 110 111 21

22. Outline Prereqs, Learning Goals, and Quiz Notes Prelude: “Additive Inverse” Problems and Discussion –Clock Arithmetic and Two’s Complement –1/3 Scottish and Fractions in Binary –Programs and Numbers –Programs as Numbers Next Lecture Notes 22

23. Problem: 1/3 Scottish Problem: Can you be 1/3 Scottish? 23

24. Can you be one-third Scottish? Mom: ?? Dad: ?? Focus on Mom (and Mom’s Mom and so on). We’ll just make Dad “Scot” or “Not” as needed at each step.

25. Can you be one-third Scottish? Mom: 2/3 Dad: Not Mom: ?? Dad: ??

26. Can you be one-third Scottish? Mom: 2/3 Dad: Not Mom: 1/3 Dad: Scot Mom: ?? Dad: ??

27. Can you be one-third Scottish? Mom: 2/3 Dad: Not Mom: 1/3 Dad: Scot Dad: Not Mom: 1/3 Dad: Scot And so on... What’s happening here? Mom: 2/3

28. Dad: Not Mom: 1/3 Dad: Scot Mom: 2/3 Dad: Not Mom: 1/3 Dad: Scot Now, focus on Dad... We can represent fractions in binary by making “Scottish family trees”: 0. 0 1 0 1...

29. Mom: 0.75 Dad: Not Mom: 0.5 Dad: Scot Mom: 0.0 Dad: Scot Here’s 0.375 in binary... 0. 0 1 1

30. Problem: 1/3 Scottish Which of the following numbers can be precisely represented with a finite number of digits/bits using a “decimal point”-style representation in base 10 but not base 2? a.1/9 b.1/8 c.1/7 d.1/6 e.None of these. 30

31. So... Computers Can’t Represent 1/3? No! Using a different scheme (e.g., in Java, a rational class with numerator and denominator fields), computers can perfectly represent 1/3! The point is: Representations that use a finite number of bits (all of them) have weaknesses. Know those weaknesses and their impact on your computations! 31

32. Outline Prereqs, Learning Goals, and Quiz Notes Prelude: “Additive Inverse” Problems and Discussion –Clock Arithmetic and Two’s Complement –1/3 Scottish and Fractions in Binary –Programs and Numbers –Programs as Numbers Next Lecture Notes 32

33. What Doesn’t Work is Not Always Obvious (1 of 2) Class Main { public static void main(String[] args) { // Let's add up 4 quarters. System.out.println("4 quarters gives us:"); System.out.println(0.25 + 0.25 + 0.25 + 0.25); // Let's do something a hundred times. int i = 100; do { // Make i one smaller. i--; } while (i > 0); System.out.println("Done!"); System.out.println("i ended up with the value: " + i); System.out.println("It went down by: " + (100 - i)); } 33

34. What Doesn’t Work is Not Always Obvious (2 of 2) Class Main { public static void main(String[] args) { // Let's add up 10 dimes. System.out.println("10 dimes gives us:"); System.out.println(0.1 + 0.1 + 0.1 + 0.1 + 0.1 + 0.1 + 0.1 + 0.1 + 0.1 + 0.1); // Let's try do something a hundred times.. // but accidentally go forever int i = 100; do { // Make i one LARGER. Oops! i++; } while (i > 0); System.out.println("Done!"); System.out.println("i ended up with the value: " + i); System.out.println("It went down by: " + (100 - i)); } 34

35. Number Representation Prediction // Let's add up 10 dimes. System.out.println("10 dimes gives us:"); System.out.println(0.1 + 0.1 + 0.1 + 0.1 + 0.1 + 0.1 + 0.1 + 0.1 + 0.1 + 0.1); // Let's try do something a hundred times.. // but accidentally go forever int i = 100; do { // Make i one LARGER. Oops! i++; } while (i > 0); 35 Can 0.1 be represented in base 2? Will i > 0 always be true? a.No and........................................ no. b.No and........................................ yes. c.Yes and....................................... no. d.Yes and....................................... yes. e.None of these (b/c the answers are not so straightforward)

36. Outline Prereqs, Learning Goals, and Quiz Notes Prelude: “Additive Inverse” Problems and Discussion –Clock Arithmetic and Two’s Complement –1/3 Scottish and Fractions in Binary –Programs and Numbers –Programs as Numbers Next Lecture Notes 36

37. Problem: Java Byte Code Problem: When compiled to bytecode, i = 100 might be “push 100; store in variable 1”. The “opcode” for bipush (push a byte) is 16 10. The opcode for istore_1 is 60 10. Here’s a typical “hex” view of ~1/5 th of the previous program’s byte code. Where is i = 100 ? 37 a d b c e: None of these. b

38. Problem: Binary Byte Code Why would the same task (finding a particular snippet of code in a bytecode file) be much more difficult if the file were represented in binary? a.Because we would have to translate all the opcodes and values to binary. b.Because many bytecode files would have no binary representation. c.Because the binary representation of the file would be much longer. d.Because data like 1100100 (100 in base 2) might not show up as the sequence of numbers 1 1 0 0 1 0 0. e.It wouldn’t be much more difficult. 38

39. Problem: Decimal Byte Code Why would the same task (finding a particular snippet of code in a bytecode file) be much more difficult if the file were represented in decimal? a.Because we would have to translate all the opcodes and values to decimal. b.Because many bytecode files would have no decimal representation. c.Because the decimal representation of the file would be much longer. d.Because data like 100 might not show up as the sequence of numbers 1 0 0. e.It wouldn’t be much more difficult. 39

40. Outline Prereqs, Learning Goals, and Quiz Notes Prelude: “Additive Inverse” Problems and Discussion –Clock Arithmetic and Two’s Complement –1/3 Scottish and Fractions in Binary –Programs and Numbers –Programs as Numbers Next Lecture Notes 40

41. Learning Goals: In-Class By the end of this unit, you should be able to: –Critique the choice of a digital representation scheme—including describing its strengths, weaknesses, and flaws (such as imprecise representation or overflow) —for a given type of data and purpose, such as (1) fixed-width binary numbers using a two’s complement scheme for signed integer arithmetic in computers or (2) hexadecimal for human inspection of raw binary data. 41

42. Next Lecture Learning Goals: Pre-Class By the start of class, you should be able to: –Use truth tables to establish or refute the validity of a rule of inference. –Given a rule of inference and propositional logic statements that correspond to the rule’s premises, apply the rule to infer a new statement implied by the original statements. 42

43. Next Lecture Prerequisites Read Section 1.3. Solve problems like Exercise Set 1.3, #1, 3, 4, 6-32, 36-44. Of these, we’re especially concerned about problems like 12-13 and 39- 44. Many of these problems go beyond the pre-class learning goals into the in-class goals, but they’re the tightest fit in the text. Complete the open-book, untimed quiz on Vista that is due before next class. 43

44. snick  snack Some Things to Try... (on your own if you have time, not required) 44

45. Problem: Weighty Numbers Problem: You have a balance scale and four weights. You may choose the mass of the weights, as long as they’re in whole units of grams. What’s the largest number n such that you can exactly measure every weight 0…n ? 45 ?g

46. 46 Problem: Representing Data Problem: Devise two different ways to represent each of the following with bits: black-and-white images text the shape of your face

47. 47 Representing Characters

48. Problem: 256-hour Clock Arithmetic Problem: Imagine you’ve built a computer that uses 256-hour clock faces, each with a single dial, as storage units. How would you store, add, subtract, and negate integers? 48

49. Concept Q: 256-hour Clock Arithmetic Java has a type called “byte” that is an 8-bit signed integer. What will the following code print? byte b = 70; b = b + 64; a.A positive number, greater than 70. b.70. c.A positive number, less than 70. d.0. e.A negative number. 49

50. Problem: Number Rep Breakdown Problem: Explain what’s happening in each of these… 50

51. Program for Introductions: Testing Java version with 100: 9900 Java version with 8675309: 265642364 Java version with 1526097757: -645820308 Reminder: this is to calculate n*(n-1), the number of introductions for a group of size n. 51

52. Program for Introductions: Testing Haskell version with 100: 9900 Haskell version with 8675309: 75260977570172 Haskell version with 1526097757: 2328974362394333292 Reminder: this is to calculate n*(n-1), the number of introductions for a group of size n. 52