Feistel Help

I never have seen a plain feistel structure, but implementations like skipjack and DES etc... So based on theories and drawings this is my VERY simple implementation. Am I on the right track? I know I will be adding a better rounding function (F) but using an easy one for now...

Code:

typedef unsigned int  WORD;
typedef unsigned char BYTE;

inline WORD F (BYTE k) {
   return (((~k) << 8) + (k));
}

void PEA (const bool encrypt, string &d, const string k, const int begin) {
   const int ROUNDS = 32;       /* Number of Rounds */
   const int SIZE = k.size();   /* Size of key */
   WORD temp, left, right;      /* 32 Bit Words */
   int index, step;             /* For Feistel Structure */
   
   /* Change direction depending on encrypt or decrypt */
   if (encrypt) index = 0, step = 1;
   else index = ROUNDS-1, step = -1;
   
   /* Push data bytes into words */
   left =  ((d[begin] << 8) + d[begin+1]);
   right = ((d[begin+2] << 8) + d[begin+3]);
   
   /* Feistel Structure */
   for (int loop = 0; loop < ROUNDS; loop++) {
      temp = right;
      right = left ^ F(k[index%(SIZE-1)]);
      left = temp;
      index += step;
   }
   
   /* Unpack words into bytes */
   d[begin+3] = (left & 255), d[begin+2] = (left >> 8) & 255,
   d[begin+1] = (right & 255), d[begin] = (right >> 8) & 255;
}

Code:

for (int loop = 0; loop < ROUNDS; loop++) {

Was that supposed to use index and step?

I'm sorry, I don't know what "feistel structure" is, so I can't help too much.

It seems to me that you don't need the parameter begin. You would just call the function with string+begin rather than func(string, begin).

I know but it only uses 4 bytes of data at a time so i had a string that i would pass the current location through PEA.

In response to your first post, no. Because step changes depending on encrypt or decrypt.

Code:

void func1(char *s, int begin) {
    cout << (s+begin);
}

func1(string, 3);

is the same as

Code:

void func2(char *s) {
    cout << s;
}

func2((string+3));

Ok well I wanted to know if the structure was right, it doesnt matter about the begin etc...

Begin does what I want it to so ill worry about it later.

Originally Posted by dwks

Code:

void func1(char *s, int begin) {
    cout << (s+begin);
}

func1(string, 3);

is the same as

Code:

void func2(char *s) {
    cout << s;
}

func2((string+3));

Well thats not the concept either. its not s + begin, its s[begin] so you cant do the second thing...

No, you just go *s.

[edit]
Or *(s+0) or s[0] etc.
[/edit]

Well I'm not using char indexing. I'm using c++ strings. So you cant add 3 to string...

Oops, sorry, I missed that.