.LP
Standard functions recognized by the Khoros expression
parser include:
.sp
.nf
\f(CW
"sin"         sin(x)          sine of x
"cos"         cos(x)          cosine of x
"tan"         tan(x)          tangent of x

"sinh"        sinh(x)         hyperbolic sine of x
"cosh"        cosh(x)         hyperbolic cosine of x
"tanh"        tanh(x)         hyperbolic tangent of x

"asin"        asin(x)         arc sine of x (range -J/2 to J/2)
"acos"        acos(x)         arc cosine of x (range 0 to J)
"atan"        atan(x)         arc tangent of x (range -J/2 to J/2)
"atan2"       atan(y,x)       arc tangent of y/x (range -pi to pi)

"sqrt"        sqrt(x)         square root of x
"hypot"       hypot(x,y)      Euclidean distance (returns  sqrt(x*x + y*y)) 
"exp"         exp(x)          exponential function of x
"expm1"       expm1(x)        (exp(x) - 1) even for small x
"ln"          ln(x)           natural log of x
"log10"       log10(x)        log base 10 of x
"log1p"       log1p(x)        log(1+x) even for small x
"gamma"       gamma(x)        log gamma function of x

"abs"         abs(x)          absolute value of |x|
"floor"       floor(x)        largest integer not greater than x
"ceil"        ceil(x)         smallest integer not less than x
"rint"        rint(x)         

"pow"         pow(x,y)        x raised to the power of y
"fact"        fact(x)         integer factorial of x
"gnoise"      gnoise(x)       gaussian noise function
"unoise"      unoise(x)       uniform noise function
"impulse"     impulse(x)      impulse function
"step"        step(x)         unit step function
"sign"        sign(x)         sign function

"j0"          j0(x)           bessel functions 
"j1"          j1(x)           bessel functions
"y0"          y0(x)           bessel functions
"y1"          y1(x)           bessel functions

"erfc"        erfc(x)         error function for 1.0 - erf(x)
"erf"         erf(x)          error function of x
     where erf(x) = 2/sqrt(pi)*integral from 0 to x of exp(-t*t) dt.
.\fP
.fi
.SH 
INSTRUCTION or STATEMENTS
.LP
Currently there is only one type of instruction, an "IF/ELSE" instruction.
An if/else expression can be entered two ways.  The following example sets
"i" performs a modulo increment of the variable i, (i = (i+1)%11).
.IP 1)
The first is the standard C style if statement.
.br
      i = if (i > 10) 0 else i + 1
.IP 2)
The second is an internal C style if statement.
.br
i = (i > 10) ? 0 : i + 1
.SH
USER DEFINED FUNCTIONS
.LP
Users may also define there own function similar to defining variables.  In
some instances this proves to be easier than trying to keep track of many
dependent expressions or evaluating complex expressions.  For instance:
.IP
f(x,y)= cos(sqrt(x**2+y**2))*exp(0.0-sqrt(x**2+y**2)/4)
.LP
To examine the above expression at x = 0 & y = 0 you simple enter f(0,0) which
should result in the value of "1.0".  You may also enter expressions as well
as variables in evaluating the function "f".  Meaning that it is valid to
enter f(-pi*2,2+x) to be evaluated.  Also, since x & y are defined by the
function then any local definition of these variable will be ignored.  Which
also implies that if you enter an expression that is dependent on a variable
not defined within the function then the current variable list will be
searched.
.IP
f(x) = x + z
.LP
The variable z is not defined by the function "f" so the current variable list
will be searched (if the variable is not defined then an error will be
displayed).
.PP
Recursive definitions are allowed but be careful to understand that there are
no checks on the stack size, therefore it is very easy to core dump the
current application.  In any recursive function you need to make sure that
the recursive expression has a base case, thus preventing a stack overflow
from occurring.  The proper way to do a Fibonacci sequence is to enter the
following expression:
.IP
fib(n) = (n == 0 || n == 1) ? n : fib(n-1) + fib(n-2)
.LP
Therefore the Fibonacci sequence of 10 "fib(10)" should return 55.  The base 
case occurs when n equals 0 or 1.  But if the user enters fib(-2) this will
cause an stack overflow.  Therefore,
.IP
fib(n) = (n > 1) ? fib(n-1) + fib(n-2) : n
.LP
should prevent the application from crashing, although this is not the
strict definition of a Fibonacci definition.
