beginning c++

beginning c++

• Michal Malecki - MSFT

  You can find rand() in the Visual Studio/MSDN documentation.  Not only is rand() not a "great" psuedo-random number generator as PRNG's go, but Martin is correct about the problems of using mod (and I didn't appreciate how bad it was until you gave the A-P statistic).  I think that might be more a result of using a non-power-of-2 modulus.
  
  There are creative ways to use up all 15 bits that rand() sets in order to make the distribution more even.  You could for instance, "blend" the bits, 3 groups of 5 bits each--this will be a 5-bit quantity in the range 0-31.  Keep repeating this until you get a number in the 0-25 range.  (This way we aren't mapping 26-31 to any of the 26 letters, causing unevenness.).
  
  assert( RAND_MAX == 32767 );  // just in case i'm wrong about RAND_MAX
  int result;
  do
  {
     int n = rand();
     result = ((n >> 10) & 0x1f) ^ ((n >> 5) & 0x1f) ^ (n & 0x1f);
  } while( result >= 26 );
  char c = (char)result + 'A';
  
  For the various operators I've used here, you should be able to find them in the documentation, and of course in any C++ book.  (Why wait until it's covered in class :))
  
  I ran this 1,000,000 times and tallied the frequencies, and the results look fairly even.  1/26 = 0.0384615...
  
  65 38554 0.038554
  66 38412 0.038412
  67 38527 0.038527
  68 38699 0.038699
  69 38084 0.038084
  70 38426 0.038426
  71 38206 0.038206
  72 38662 0.038662
  73 38577 0.038577
  74 38318 0.038318
  75 38511 0.038511
  76 38360 0.03836
  77 38189 0.038189
  78 38494 0.038494
  79 38646 0.038646
  80 38725 0.038725
  81 38516 0.038516
  82 38069 0.038069
  83 38849 0.038849
  84 38742 0.038742
  85 38276 0.038276
  86 38295 0.038295
  87 38511 0.038511
  88 38395 0.038395
  89 38684 0.038684
  90 38273 0.038273
  
  Brian
  
  

• caaptain

  Thanks, I was ready to get into a fist fight with this thing. I've only been taking C++ for about 6 weeks now, and in that lab for school, my book only vaguely described the usage of the random function. This is the short program I was working on, once again, thanks for your help!
  

  #include <iostream.h>
  #include <stdlib.h>   //for random 
  #include <iomanip.h>  //for setw

  void get_a_letter(int);
  int main()
  {
   int seed;
   char answer='y';
      while(answer=='y')
      {
      cout<< "\n Please enter integer value for seed.";
      cin>>seed;
      srand(seed);
      get_a_letter(seed);
      cout<< "\n Would you like to see more random letters ";
      cin>> answer;
      }
   cout<<endl;
   return 0;
  }

  void get_a_letter(int seed)
  {
      int letter_ctr, rand_num, abc;
      for(letter_ctr=0; letter_ctr<15; ++letter_ctr)
      {
       rand_num=rand()%26+65;
       abc=char(rand_num);
       cout << "\n " << letter_ctr+1<< setw(5) << (char)abc;
      }
  }

  
  

• Ofer Gal

  Just to know:
  The modulo operator does its work but it has the drawback that the sequences of random numbers are not evenly distributed.
  This is no problem if the range you need is far less than RAND_MAX.
  Example: In the case the range is is from 0 to RAND_MAX*2/3 you get the problem that the lower 1/3 range of numbers are choosen twice as often as the upper 2/3.
  
  

• Anwar Javed

  You can the % (modulo) operator to constrain the result of rand() to a range (0 through 25 = 26 values), and then add an offset (65).
  
  int randnum = rand() % 26 + 65;
  char mychar = (char)randnum;
  
  

  
  

• Jmill07

  An easy way to make sure each integer in your set has the same number of rand() results mapped to each is to just dissallow results from the top incomplete set:

  const int LO = 65, HI = 90;
  const int MY_RAND_MAX = RAND_MAX - (RAND_MAX + 1) % (HI - LO + 1);
  
  int randomInt;
  
  do {
    randomInt = rand();
  } while(randomInt > MY_RAND_MAX);
  
  randomInt = randomInt % (HI - LO + 1) + LO;
  
  Really, the same thing Brian is doing.

  
  

• jcbrooks75

  @wooten
  Beside all problems and the evency of rand, the main problem is the modulo operation itself. rand gives a random nimber from 0 up to 0x7fff. So it uses 15bits (hopefully evenly distributed).
  The module is nothing else then the left over of a division. So the modulo breaks the hole sequence of numbers into sub sequences from A to Z.
  So starting at
  0-25 its the first sequence of A-Z
  26-51 is the second sequence
  52-77 is the third sequence
  this continus at a total 1260 times up to the
  32760-32767

  You see that there is no longer a full seqence. So if the modulo is not taken from a power of 2 number the random number seuqnece is not evenly distributed.
  And you can see that the lower 7 chars (A,B,C,D,E,F,G) choosen 1/1260 more often.
  
  HTH

  
  
  

• Jaewon

  That would explain why I would get the lower letters of the alphabet more often than the upper letters. I did find it kind of odd that I was getting more of the letters from A-P and very seldomly getting the rest of the alphabet. Could you explain this function and how to use it in the program I posted above As I mentioned before, I'm beginning C++ and have only been in the intro to c++ class for about 6 or 7 weeks now and the modulo and random functions are extremely new and were very vaguely described to me in the text books or class. I have more experience with VB than C++ which I very quickly learned are about as different as day and night. I appreciate your time, patience, help and tips on C++ and explainations of the different functions and operators.
  
• beginning c++