randomizer

[Josh_Dude83]

I've been looking around and trying to get a random number generator easy right? Not really I guess. Here is what I have but I'm getting an error "undefined reference to randomize"

int deal(void)
{
    int random;
    randomize();
    random = rand() % 11;
    
    return random;
}

Also have a question about strcmp, can you compare against a string like this:

while(strcmp(choice, 'quit') == 0);

[itsme86]

1) The function you're looking for is srand(). Hint: You can use time(0) as an adequate seed in most cases.

2) You need to use double quotes, not single.

[msh]

Minor nit-pick. I believe it's time( NULL ) for C, and time( 0 ) for C++. Or does it make no difference?

[grumpy]

Minor nit-pick. I believe it's time( NULL ) for C, and time( 0 ) for C++. Or does it make no difference?

No difference. In both C and C++, a pointer with value zero is equivalent to the NULL pointer.

[Prelude]

>randomize();
>random = rand() % 11;
Assuming you replace randomize with srand, it's still not going to work properly. The generator should be seeded very rarely, if more than once. Seeding every time you call rand will produce rather nonrandom numbers.

[PehJota]

Additional little nit pick: using the modulo operator to shrink the range of rand() is okay for a simple program where uniformity is of no concern, but the results of such randomization are skewed. Using a pseudorandom deviate (as explained in this informative page) is a better solution. Basically, you produce a pseudorandom deviate between 0 and 1 by calling rand() and dividing it by the number of different results it offers (one more than RAND_MAX), then you multiply that by the number of results you want.

int deal(void)
{
    int random;
    random = rand() * (1.0 / (RAND_MAX + 1.0)) * 11;
    return random;
}

[msh]

>randomize();
>random = rand() % 11;
Assuming you replace randomize with srand, it's still not going to work properly. The generator should be seeded very rarely, if more than once. Seeding every time you call rand will produce rather nonrandom numbers.

I did some (very) basic testing on this, and if the application runs at user-speed, e.g. there is a delay for user input between each srand() call, the results are satisfactory. If the application runs at computer-speeds, however, it's pretty much broken.

[Prelude]

>If the application runs at computer-speeds, however, it's pretty much broken.
The time function works with second granularity. If your user interface is clunky enough that users are physically unable to make a request without waiting at least one second, calling srand immediately before every call to rand should work.

However, I still see no reason why you would want to do it this way. The period of the random number generator should be more than enough to satisfy your randomization needs with a single call to srand. Further, you'd be added unnecessary overhead by calling srand when it doesn't need to be called.