I loved this question. I have just spent a few months reviewing this exact thing and how the data space is handled for different data types and nullable/non-nullable columns. As well as how clustered indexes (PK) become fragmented and at what rate based on a continually increasing PK.