Re: No Go To's Forever!

[tmoran@bix.com]

> look_ahead(in_file,ch,bool);
> <<start>>
> while bool loop
>   skipline(infile);
>   look_ahead(in_file,ch,bool);
> end loop;
>
> if ch=' ' then
>   while ch=' ' loop
>     get(infile,ch);
>     look_ahead(in_file,ch,bool);
>   end loop;
>   goto start;
> end if;

How about:

loop
  look_ahead(in_file,ch,got_eol);
  if got_eol then
    skipline(infile);
  elsif ch = ' '
    get(infile, ch);
  else
    exit;
end loop;

It's shorter, the logic of what's going on is clearer, only
a single line needs to be changed to add tab skipping,
there's only one way to enter the code sequence (no distant
"goto start") so the first reference to bool or ch is
guaranteed to come *after* the look_ahead call that sets
their values, and, since the compiler can tell the control
flow, it can do better optimization.

>I realize everything is posible to do without gotos
>but I canot figure out why it is so horible.
   A piece of logic expressed with loops, ifs, cases,
and subroutine calls is almost always clearer (and
therefore less likely to be erroneous) and is commonly
also shorter than the (attempted) equivalent using goto.

In MSDOS .bat and assembler you have no choice but to use
a jump/goto.  Your toolset is bigger now, learn to use it.

>I know it might make the program hard to follow, but I don't care!
  But a hard to follow program is hard to debug and hard to
extend, and therefore worth less.