### 4Mathematical Functions

This chapter describes the basic mathematical constants and functions provided by the PLT Scheme Science Collection.

The constants and functions described in this chapter are defined in the "math.ss" file in the science collection and are made available using the form:

 (require (planet williams/science/math))

#### 4.1Mathematical Constants

The following table shows the mathematical constants defined by the PLT Scheme Science Collection:

 e The base of the exponentials, e log2e The base two logarithm of e, log2 e log10e The base ten logarithm of e, log10 e sqrt2 The square root of two, √2 sqrt1/2 The square root of one half, √½ sqrt3 The square root of three, √3 pi The constant pi, π 2*pi Pi time two, 2π 4*pi Pi times four, 4π pi/2 Pi divided by two, π/2 pi/4 Pi divided by four, π/4 sqrtpi The square root of pi, √π 2/sqrtpi Two divided by the square root of pi, 2/√π 1/pi The reciprocal of pi, 1/π 2/pi Twice the reciprocal of pi, 2/π ln10 The natural log of ten, ln 10 or loge 10 1/ln10 The inverse of the natural log of ten, 1/ln 10 ln2 The natural log of two, ln 2 or loge 2 1/ln2 The inverse of the natural log of two, 1/ln 2 lnpi The natural log of pi, ln π or loge π euler Euler’s constant, γ

#### 4.2Testing for Infinities and Not-a-Number

PLT Scheme provides +inf.0 (positive infinity), -inf.0 (negative infinity), +nan.0 (not a number), and +nan.0 (same as +nan.0) as inexact numerical constants. The following functions are provided as a convenience for checking for infinities and not-a-number.

 (nan? x) → boolean? x : any/c
Returns true, #t, if x is not-a-number (i.e., equivalent to +nan.0 or, equivalently, +nan.0), and false, #f, otherwise. Note that this is not the same as (not (number? x)), which is true if x is not a number.

 (infinite? x) → boolean? x : any/c
Returns 1 if x is positive infinity (i.e., equivalent to +inf.0), -1 if x is negative infinity (i.e., equivalent to -inf.0), and false, #f, otherwise. Note that (finite? x) is not equivalent to (not (infinite? x)), since both finite? and infinite? return false, #f for anything that is not a real number.

 (finite? x) → boolean? x : any/c
Returns true, #t, if x is a finite real number and false, #f, otherwise. Note that (finite? x) is not equivalent to (not (infinite? x)), since both finite? and infinite? return false, #f for anything that is not a real number.

#### 4.3Elementary Functions

The following functions provide some elementary mathematical functions that are not provide by PLT Scheme.

 (log1p x) → number? x : real?
Computes the value of log(1 + x) in a way that is accurate for small x.

 (expm1 x) → real? x : real?
Computes the value of exp(x - 1) in a way that is accurate for small x.

 (hypot x y) → real? x : real? y : real?
Computes the value of (x2 + y2)½ in a way that avoids overflow.

 (acosh x) → real? x : real?
Computes the value of the hyperbolic arccosine, arccosh, of x.

 (asinh x) → real? x : real?
Computes the value of the hyperbolic arcsine, arcsinh, of x.

 (atahh x) → real? x : real?
Computes the value of the hyperbolic arctangent, arctanh, of x.

 (ldexp x e) → real? x : real? e : integer?
Computes the value of x × 2e.

(frexp x)
 real? integer?
x : real?
Splits the real number x into a normalized fraction f and exponent e such that x = f × 2e and 0.5 ≤ f < 1. The function returns f and e as multiple values. If x is zero, both f and e are returned as zero.

#### 4.4Testing the Sign of Numbers

 (sign x) → (integer-in -1 1) x : real?
Returns the sign of x as 1 if x ≥ 0 and -1 if x < 0. Note that the sign of zero is positive, regardless of its floating-point sign bit.

#### 4.5Approximate Comparisons of Real Numbers

It is sometimes useful to be able to compare two real (in particular, floating-point) numbers approximately to allow for rounding and truncation errors. The following functions implements the approximate floating-point comparison algorithm proposed by D.E. Knuth in Section 4.2.2 of Seminumerical Algorithms (3rd edition) [Knuth].

 (fcmp x y epsilon) → (integer-in -1 1) x : real? y : real? epsilon : real?
Determines whether x and y are approximately equal to within a relative accuracy, epsilon. The relative accuracy is measured using an interval of 2 × delta, where delta = 2k × epsilon and k is the maximum base 2 exponent of x and y as computed by the function frexp. If x and y lie within this interval, they are considered equal and the function returns 0. Otherwise, if x < y, the function returns -1, or if x > y>, the function returns 1.

The implementation of this function is based on the packege fcmp by T.C. Belding.

#### 4.6Log10 and Decibels (dB)

 (log10 x) → real? x : real?
Returns the log base 10 of x, log10(x).

 (dB x) → real? x : real?
Returns the value of x in decibels, 10*log10(x).

#### 4.7Assure Floats for Unsafe Code

These routines are used to assure that a real variable (or vector) is represented as a floating-point number. This can be used to protect code that uses the new unsafe floating-point operations.

 (real->float x) → inexact-real? x : real?
Returns an inexact real (i.e., a float) given a real x. Raises an error if x is not a real. This can be used to insure a real value is a float, even in unsafe code.

 (real-vector->float-vector v) → (vector-of inexact-real?) v : (vector-of real?)
Returns a vector of inexact reals (i.e., floats) given a vector of reals, v. Raises an error if any element of v is not a real.

 (with-float (x ...) expr ...)
Executes the exprs with each x guaranteed to the a float. All of the xs must be identifiers.