picasso: wrote in message to alt.comp.lang.learn.c-c++ 

> > Can anyone see any fault in this now that i should fix?

Andrew Koenig replied:

The program is much too large.

Moreover, the problem is not well specified.  Here's why I say so:  When a 
problem specification uses the word "the," doing so often indicates a 
potential trouble spot.  In this case, for example, you said you want to 
determine the biggest value in the input, and that use of "the" implicitly 
says that such a value exists!  But what if there is no input?  Then there 
is no biggest value.

Similarly, if the input contains only one value, there is no second biggest 
value either.

So the use of "the" in the problem specification has pointed out that there 
will be trouble unless there are two or more values in the input.

This observation suggests a strategy that makes it unnecessary to use magic 
numbers such as -2147483648.

For example, you might start by reading the first two elements of the input 
into the variables "biggest" and "secondbiggest."  Then, if 
secondbiggest>biggest, swap their values.

At this point, you have established an invariant:

    "biggest" contains the biggest value you've seen in the input so far.
    "secondbiggest" contains the second biggest value you've seen.

with the proviso that the two variables might be equal if the first two 
input values are equal.

Now, all you have to do is read the rest of the input, making sure to 
maintain the invariant each time you obtain a new input value.  To see how 
to do it, let's call that value x.  Now there are three cases:

    1) x <= secondbiggest.  In this case, biggest and secondbiggest are 
still the winners, so we don't need to do anything.

    2) x > biggest.  In this case, biggest is now the second biggest value 
you've seen and x is the biggest, so you have to adjust the variables 
accordingly:

            secondbiggest = biggest;
            biggest = x;

    3) All other cases.  Here, x <= biggest and x > secondbiggest, so there 
is a new second biggest value but the biggest value doesn't change.

So each trip through the loop does the following:

    if (x > biggest) {
        secondbiggest = biggest;
        biggest = x;
    } else if (x > secondbiggest) {
        secondbiggest = x;
    }

When you're done with the loop, biggest and secondbiggest will have the 
right values.  No magic necessary.

----------------------
So although it works as is, this program needs some improvement.