Comments to the Turing machine implementation (TuringMachine.cpp) by Hector Zenil (despite a couple of C++ functions, the implementation is mostly in C language):

This code initializes the Turing machine empty tape:

while(n!=step_limit){tape[n]=0;n++;}

and this initializes the transition table:
while(k<2*states) {while(j<5) {transition[k][j]=atoi(argv[n+2]);
j++;n++;}j = 0;k++;}

The core code for the operation of the Turing machine is encoding here:

head_pos=start_cel-1;inf_bound=start_cel-1;sup_bound=start_cel-1;
j=init_state;while(state!=0&&step<step_limit) { while(s<symbols){
while(j<states+1){while(k<2*states){if(transition[k][0]==j&&
transition[k][1]==s) if(state == transition[k][0] && tape[head_pos]==
transition[k][1]&&applied_flag==0){state=transition[k][2];
tape[head_pos]=transition[k][3]; head_pos=head_pos+transition[k][4];
applied_flag=1;}k++;}k=0;j++;}s++;j=1;}s=0;step++;applied_flag=0;

which is quite short.

This is the code to output the tape content on the screen:

if(state==0) for(n=inf_bound;n<sup_bound+1;n++) cout<<tape[n];
else cout<<symbols+1;cout<<"\n"<<step<<endl;

One can certainly write a program using pointers in order to avoid the initializations of the tape and the transition table making it a little bit shorter. But we are expecting people trying to follow other universal models of computation to implement a shorter version even though it has to input/output as a Turing machine (so one can ask about the question of the encoding program-size together with the implementation itself).