ok we all knw wht a merge sort is right, the code is easy too

but the problem is

how to:

Design a modified Mergesort algorithm called Mergesort3 that sorts an array by dividing it
into three equal (or almost equal) sub-arrays and then merging the three sorted sub-arrays.....????

If the code is so easy for mergesort, then why is an extra partition so difficult?
If mergesort was written properly, I imagine that you would get > 80% code reuse on your new implementation.
What specifically is the hangup?