Wrong Encryption for some Words/C - c

Question

I'm trying to make a simple encryption program in C. My aim is to translate "jim" (it can be any word) to 10,9,13. Then plus 1,2,3(because word has 3 letter) and get 11,12,16 then again translate to text, then write on screen klp.I have 2 algorithm one for encryption and second one for return first one.It s working but for these words there s problem
"compiler you executable through" when i start first algorithm and then second algortihm(inverse) i took ."compileo vou euecrqable throrgh "i checked algorithm several times i dont know where i overlooked
PS:if u want to run make txt file which named unencrypted.txt and write word inside then first algorithm then second one
int main() {
int g,z,o,c,l,i,j,k,*D;
char alfabe[26]={'a','b','c','d','e','f','g','h','i','j','k','l','m','n','o','p','q','r','s','t','u','v','w','x','y','z'};
FILE *fp1;
FILE *mat;
char word[20];
fp1 = fopen("unencrypted.txt","r+");
do {
g = fscanf(fp1,"%s",word);
if (g != EOF) {
mat=fopen("encrypted.txt","a+");
c=strlen(word);
D=(int*)calloc(c,sizeof(int));
for(i=0;i<c;i++) {
for(j=0;j<26;j++) {
if(word[i]==alfabe[j]) {
D[i]=(((j+1)+(i+1))%26);
break;
}
}
}
}
for(z=0;z<c;z++){
o=D[z];
word[z]=alfabe[o-1];
}
for(k=0;k<c;k++) {
fprintf(mat,"%c",word[k]);
}
fprintf(mat," ");
fclose(mat);
} while (g != EOF);
fclose(fp1);
}
int main() {
int g,z,o,c,l,i,j,k,*D;
char alfabe[26]={'a','b','c','d','e','f','g','h','i','j','k','l','m','n','o','p','q','r','s','t','u','v','w','x','y','z'};
FILE *fp1;
FILE *mat;
char word[20];
fp1 = fopen("encrypted.txt","a+");
do {
g = fscanf(fp1,"%s",word);
if (g != EOF) {
mat=fopen("unencrypted1.txt","a+") ;
c=strlen(word);
D=(int*)calloc(c,sizeof(int));
for(i=0;i<c;i++) {
for(j=0;j<26;j++) {
if(word[i]==alfabe[j]) {
if(0>((j+1)-(i+1))){
D[i]=((j+1)-(i+1)+26);
} else D[i]=((j+1)-(i+1));
break;
}
}
}
}
for(z=0;z<c;z++){
o=D[z];
word[z]=alfabe[o-1];
}
for(k=0;k<c;k++) {
fprintf(mat,"%c",word[k]);
}
fprintf(mat," ");
fclose(mat);
} while (g != EOF);
fclose(fp1);
}

Answer 1

First, in your encryption routine, you can have undefined behaviour:
word[z]=alfabe[o-1];
accesses memory it shouldn't if o == 0, which will happen if the sum of 1-based index of the character in the word and 1-based index in the alphabet is a multiple of 26.
To avoid that, you can set word[z] = alfabe[(o+25)%26], or, simpler, leave off one of the +1s in the previous loop and use alfabe[o].
The access of alfabe[-1], if it happens, is the sort of undefined behaviour that is unlikely to cause a segmentation fault or similar drastic things, it will probably only cause an unexpected byte to be written in the encrypted word.
In the decryption, you have similar problems,
if(0>((j+1)-(i+1))){
D[i]=((j+1)-(i+1)+26);
} else D[i]=((j+1)-(i+1));
If the word is long enough, even j - i + 26 may be negative (that is, however, extremely unlikely for natural English text - also in a lot of other languages).
And again, it can happen that D[i] is set to 0, in which case
for(z=0;z<c;z++){
o=D[z];
word[z]=alfabe[o-1];
}
will again access alfabe[-1] and cause undefined behaviour.
Note also that if your input contains anything except lower case letters and whitespace (which is discarded by the fscanf), the corresponding entry of the D array contains whatever bytes occurred to be at that memory location, again causing undefined behaviour.
If your input contains only lower case letters (and whitespace), you can find the index in the alphabet much easier than with the loop by setting
j = word[i] - 'a';
To avoid the alfabe[-1] issues, I recommend using 0-based indices.