Final answer:
The fraction of total space taken up by overhead in the binary tree implementation is (2n + 5) / (7n), where 'n' is the number of leaf nodes.
Step-by-step explanation:
To calculate the overhead for the binary tree implementation, we need to consider the space occupied by the pointers and the data objects. Let's assume the tree has 'n' leaf nodes. Since each internal node has two pointers and no data field, the space occupied by internal nodes would be 2 * 4 * (n - 1). The space occupied by the leaf nodes would be 28 * n. The total space occupied by the tree would be the sum of these two values.
Overhead = ((2 * 4 * (n - 1)) + (28 * n)) / (n * 28) = (8 * n - 8 + 28) / (28 * n) = (8n + 20) / (28n) = (2n + 5) / (7n)
Therefore, the fraction of total space taken up by overhead is (2n + 5) / (7n). This fraction cannot be further simplified as it is already in its lowest terms.