Summary: We examine a variety of issues pertaining to numeric values in Scheme, including the types of numbers that Scheme supports and some common numeric functions.

Computer scientists write algorithms for a variety of problems. Some types of computation, such as representation of knowledge, use symbols and lists. Others, such as the construction of Web pages, may involve the manipulation of strings (sequences of alphabetic characters). However, as you've seen with some of your initial experiments with images, a significant amount of computation involves numbers. One advantage of doing numeric computation with a programming language, like Scheme, is that you can write your own algorithms to make the computer automate repetitive tasks. As you do numeric computation in any language, you must first discover what types of numbers the language supports (some languages support only integers, some only real numbers, some combinations) and what numeric operations the language supports. Fortunately, Scheme supports many types of numbers (as you may have discovered in the first few labs) and a wide variety of operations on those numbers.

As you probably learned in secondary school, there are a variety of kinds of numbers. The most common types are the integers (numbers with no fractional component), rational numbers (numbers that can be expressed as the ratio of two integers), and real numbers (numbers that can be plotted on a number line). Some Scheme implementations, such as Script-Fu, the primary Scheme in GIMP, support only integers and real numbers. In some Scheme implementations, including the one used by default in MediaScript, other numeric types are available, such as the rational numbers (numbers that can be expressed as the ratio of two integers) and complex numbers (numbers with a possible imaginary component). Why does Script-Fu leave out some kinds of numbers? Because the implementers did not see a need for them. In fact, the standard language definition for Scheme says that an implementation of the language does not have to support all categories of numbers. Scheme provides two basic predicates that let us check whether or not a value has a particular type: integer? and real?. > (integer? 2) #t > (real? 2) #t > (integer? 2.5) #f > (real? 2.5) #t > (integer? "two") #f MediaScript uses integers to represent other kinds of values, such as images and RGB colors, so integer? will return true for them, too. (Don't worry that you haven't seen RGB colors yet; we'll get to them in a few days.) > (integer? (rgb-new 0 0 0)) #t > (integer? (image-load "/home/rebelsky/glimmer/samples/rebelsky-stalkernet.jpg")) #t > (integer? "/home/rebelsky/glimmer/samples/rebelsky-stalkernet.jpg") #f We will return to these predicates, and others, when we consider conditionals.

Within each category of numbers, Scheme distinguishes between exact numbers, which are guaranteed to be calculated and stored internally with complete accuracy (no rounding off), and approximations, also called inexact numbers, which are stored internally in a form that conserves the computer's memory and permits faster computations, but allows small inaccuracies (and occasionally ones that are not so small) to creep in. Since there's no great advantage in obtaining an answer quickly if it may be incorrect, we shall avoid using approximations in this course, except when the data for our problems are themselves obtained by inexact processes of measurement. To determine whether Scheme is representing a particular number exactly or inexactly, use one of the predicates exact? and inexact?. Real numbers are never represented exactly, and integers can be represented exactly or inexactly. You can convert between the two representations with exact->inexact and inexact->exact. > (exact? 2) #t > (exact? 2.0) #f > (inexact->exact 2.0) 2 The Scheme standard does not directly support the familiar category of natural numbers, but we can think of them as being just the same things as Scheme's exact non-negative integers.

All Scheme implementations support integers and reals. PLT Scheme, the Scheme implementation we normally use in MediaScheme, also supports rational numbers and complex numbers. PLT Scheme's support of rational numbers may mean that you get results as a ratio of two numbers, rather than as a decimal number. For example, when you divided 2 by 5, you might expect to get 0.4, as in > (/ 2 5) 0.4 In fact, you will see that result in many Scheme implementations, including in Script-Fu, GIMP's default Scheme implementation. However, since decimals are often approximated, PLT Scheme prefers rationals when it makes sense to use them. Hence, in MediaScheme, you'll see slightly different output. > (/ 2 5) 2/5 Most of the time, it won't really matter which representation Scheme uses. However, there are times that the results are a bit confusing when expressed in rational form. You may have seen these confusing results when computing average grades, as in the following. > (/ (+ 90 80 100 23 80 75) 6) 224/3 Often, it helps to put these numbers into inexact form. > (exact->inexact (/ (+ 90 80 100 23 80 75) 6)) 74.66666666666667 PLT Scheme provides two additional procedures for working with rational numbers, numerator and denominator. As you might expect, these return the numerator and denominator of a rational number. > (numerator 5/7) 5 > (denominator 5/7) 7 > (numerator 20/6) 10 > (denominator 20/6) 3 It's generally a bad idea to use these procedures with inexact numbers, as PLT Scheme chooses different values than most normal people expect. > (numerator 0.4) 3602879701896397.0 > (denominator 0.4) 9007199254740992.0 > (/ (numerator 0.4) (denominator 0.4)) 0.4

Section 6.2.5 of the Revised5 report on the algorithmic language Scheme lists Scheme's primitive procedures for numbers. Read through the list at this point to get a feel for what Scheme supports. The following notes explain some of the subtler features of commonly used numerical procedures. As you read about procedures, think about how you might use them in writing color filters or in other graphical algorithms. Warning! The output from different Scheme intpreters are inconsistent, and sometimes even inconsistent with our expectations. In a few cases, you may see slightly different responses than appear in this reading. As you've already seen, the addition and multiplication procedures, + and *, accept any number of arguments. You can, for instance, ask Scheme to imitate a cash register with a command like this one: > (+ 1.19 .43 .43 2.59 .89 1.39 5.19 .34 ) 12.45 You can call the - procedure or the / procedure to operate on a single argument. The - procedure returns the additive inverse of a single argument (its negative), the result of subtracting it from 0. The max procedure returns the largest of its parameters and the min procedure returns the smallest of its parameters. As we've already seen, max can be useful when you want to ensure that a computation returns a value no smaller than a certain value and min can be useful when you want to ensure that a computation returns a value no larger than a desired maximum value.

There are four procedures that relate to division (/, quotient, remainder, and modulo). You've already seen that / can divide one value by another. If you call the / procedure with a single parameter, it returns the multiplicative inverse of that parameter (its reciprocal), the result of dividing 1 by it. > (/ 2) 0.5 > (/ 1) 1 > (/ 0.5) 2 > (/ 0) /: division by zero The quotient and remainder procedures apply only to integers and perform the kind of division you learned in elementary school, in which the quotient and the remainder are separated: Thirteen divided by four is three with a remainder of one: > (quotient 13 4) 3 > (remainder 13 4) 1 > (quotient 1 2.5) quotient: expects type <integer> as 2nd argument, given: 2.5; other arguments were: 1 As the final example suggests, quotient can only be applied to integers. The / procedure, on the other hand, can be applied to numbers of any kind (except that you can't use zero as a divisor) and yields a single result. The modulo procedure is like remainder, except that it always yields a result that has the same sign as the divisor. In particular, this means that when the divisor is positive and the dividend is negative, modulo yields a positive (or zero) result. (When can a remainder be negative? Consider -7 divided by 3. Do we think of -7 as -2*3-1 or -3*3+2? Scheme makes the former decision for remainder and the latter decision for modulo.) > (remainder -13 4) -1 > (modulo -13 4) 3 > (remainder 13 -4) 1 > (modulo 13 -4) -3 The modulo procedure can be particularly useful when you want to ensure that a value falls in a certain range, and you don't just want higher values to map to the highest value in the range. For example, you'll find many times this semester that you want to compute a number between 0 and 255, but end up computing something out of that range. we can ensure that they fall within the appropriate range with max and min. We can get somewhat different effects by using (modulo computed-value 256). This expression ensures that the value is between 0 and 255, but causes larger numbers to wrap-around to become smaller numbers. > (define blue-component 250) > (min 255 (+ 32 blue-component)) 255 > (modulo (+ 32 blue-component) 256) 26 You can also think of the value of (modulo value modulus) as follows: We break the number line up into modulus-sized sections and then find the offset of value from the start of that section. For example, if we use a modulus of 10, the non-negative sections of the number line would be (0..9), (10..19), (20..29), and so on and so forth. The number 23 would be in the section (20..29). Since it's 3 bigger than 20, (moduo 23 10) is 3.

At times, we will have a real number and will want to convert it to a nearby integer. For example, if you are working with images, the components of an RGB should be integers; weird things can happen if you try to use real numbers (not always, but sometimes). Similarly, when specifying a row and a column in an image, we want whole numbers for those row and column. Scheme provides four basic procedures for this conversion: round, truncate, floor, and ceiling. You will explore the differences between these procedures in the corresponding lab. Warning! At times, the Scheme interpreter will complain that it is expecting an integer but sees a real value, even when you think you have an integer. The problem is not with you, but with the error messages. Most of the time that the interpreter says that it wants an integer, it really wants an exact integer, so use inexact->exact to get the number in the correct form.

Scheme provides five basic predicates for comparing numeric values, < (less than), <= (less than or equal to), = (equal to), >= (greater than or equal to), and > (greater than). When given two arguments, they return #t if the indicated relation holds between the two arguments. > (< 5 10) #t > (> 5 10) #f These predicates can also take more than two arguments. Each predicate returns #t only if the relation holds between each pair of adjacent arguments. If the relation fails to hold between a pair of adjacent arguments, the predicate returns #f. > (< 2 3 4) #t > (< 2 3 1) #f The log procedure, despite its name, computes natural (base e) logarithms rather than common (base ten) logarithms. You can convert a natural logarithm into a common logarithm by dividing it by the natural logarithm of 10. In case you've forgotten, the common logarithm of n is the power to which you raise 10 in order to get n. > (log 100) 4.605170185988092 > (/ (log 100) (log 10)) 2.0 Scheme provides the standard host of trigonometric functions, which include sin, cosine, and tan. When using these functions, remember that all angles are measured in radians, not degrees. > (sin 90) 0.8939966636005579 > (cos 90) -0.4480736161291701 > pi 3.141592654 > (exact? pi) #f > (sin (/ pi 2)) 1.0 > (cos (/ pi 2)) 6.123031769e-17 You may wonder why the cosine of pi-over-2 (a right angle) is not 0. It's because pi is not exactly the value of pi, but is rounded off. However, as scientific notation indicates, the value is pretty close to 0. (There are sixteen leading 0's.) We can use the trigonometric functions when we start doing more involved drawings (e.g., they can help us draw polygons). The trigonometric functions also provide the opportunity to do some interesting color transformations.

Many students are puzzled by both the modulo and remainder procedures. For remainder, you really should think back to middle-school math: the remainder is what's left after whole-number division. Since modulo is the same as remainder for positive numbers, you can think of it that way. More importantly, modulo provides an interesting way of counting. Most of the time you add 1, you follow standard protocols (1 plus 1 is 2, 2 plus 1 is 3, ...). However, when you reach the modulus value, you go back to zero. The following table shows the value of remainder and modulo for a variety of values.

n	-4	-3	-2	-1	0	1	2	3	4	5	6	7	8
(remainder n 3)	-1	0	-2	-1	0	1	2	0	1	2	0	1	2
(remainder n 4)	0	-3	-2	-1	0	1	2	3	0	1	2	3	0
(modulo n 3)	2	0	1	2	0	1	2	0	1	2	0	1	2
(modulo n 4)	0	1	2	3	0	1	2	3	0	1	2	3	0
(modulo n 5)	1	2	3	4	0	1	2	3	4	0	1	2	3