MB> Be careful: srand() takes an unsigned int, while time() 
 MB> returns a time_t.  These are different sizes on many 
 MB> compilers.  To handle this portably, do something like this:

 MB>    srand((unsigned int)(time(NULL) % UINT_MAX));

        Yeah, I did something like that.  Although I ran the time() result
through the ELF Hashh  first, and used INT_MAX.

 MB> If you need to guarantee uniform distribution of inputs 
 MB> to srand(), the proper solution is to apply a hash 
 MB> (such as the ELF hash I discussed in a prior message to 
 MB> you) and a prime modulus less than INT_MAX before 
 MB> feeding the seed to srand().

        Why a prime modulus?

 MB> Note especially that the random number generators 
 MB> provided with most compilers are awful.  Borland's and 
 MB> Watcom's are particularly bad, since even their 32-bit 
 MB> OS/2 compilers only return 16 bits of significance from 
 MB> rand().  ANSI C guarantees that RAND_MAX is at least 
 MB> 32767, and Watcom gives you the bare minimum.  Just 
 MB> because rand() returns a 32-bit int does not 
 MB> necessarily imply that the most significant 16 bits are 
 MB> ever non-zero.


        Interesting.  Fortunately, this application doesn't really require
a good random number generator.  The program is just a simple text based
background bitmap changer.

Sean


--- Maximus/2 3.01