On Mon, Jan 20, 2003 at 12:07:31PM -0500, David Z Maze wrote:
> [EMAIL PROTECTED] (Robert Land) writes:
> > On Fri, Dec 13, 2002 at 11:23:43AM -0500, Derrick 'dman' Hudson wrote:
> >> "imperative" and "procedural" are the same thing, and C is a prime
> >> example. It is such because the structure of a C program is a
> >> collection of procedures which start with "main". Each procedure is a
> >> linear list of statements to be executed in order.
> >
> > Could you specify a "linear list" more clearly? - the
> > contrary would be a "nonlinear list" which on the first
> > view seems to be self-contradictory.
>
> For example, in C:
>
> int fib(int n)
> {
> if (n < 2)
> return n;
> return fib(n-2) + fib(n-1);
> }
>
> the rules require that fib(n-2) is called before fib(n-1). In Scheme:
This is incorrect. There is no sequence point in the statement:
return fib(n-2) + fib(n-1);
So, it's up to the implementation which side of the '+' is evaluated
first.
> (define (fib n)
> (if (< n 2)
> n
> (+ (fib (- n 2)) (fib (- n 1)))))
The C and Scheme functions are essentially identical. In C, statements
are executed in order. I'm not too up on functional languages, but I
seem to recall they need special syntax to execute statements
sequentially.
That fib function is painfully slow via recursion and susceptible to
uncaught overflow in the C version (Scheme has big numbers built in).
#include <stdio.h>
unsigned long long
get_nth_fib (unsigned long long count)
{
unsigned long long last2 = 0ULL;
unsigned long long last = 1ULL;
unsigned long long cur = count;
unsigned long long i;
if (count > 1) {
for (i = 1; i < count; ++i) {
cur = last + last2;
if (cur < last) {
fprintf (stderr, "get_nth_fib(%llu) overflows!\n", count);
cur = 0UL;
break;
}
last2 = last;
last = cur;
}
}
return cur;
}
Less elegant, but much faster and more robust than the recursive C version.
A table look-up would be even faster up to some 'n' (say, 93 in this case).
--
echo ">gra.fcw@2ztr< eryyvZ .T pveR" | rot13 | reverse
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
