Suppose we pick a near-universal random hash function h with range {1,2,3,4} and hash the elements in {1,2,3,4,5,6,7,8,9} with it. What is the tightest upper bound on the expected number of elements that collide with h(1)? [ignoring h(1) itself]

Can you think of any assumptions you've made about email? What path does information follow as it travels from one place to another? Create a bullet point list of your newfound knowledge. The first question to answer is what causes the differences. Think about how distinct models have varying levels of complexity (or abstraction).

What comes to mind when you think of email? What occurs after an email has been sent? Make a note of everything you have learned up to this point. Does anybody know what these disparities' root causes are, and if so, what they are? Do your models support various degrees of complexity?

When you think about email, what images pop into your head? What occurs after an email is sent? Jot down anything you can think of at the moment. Is there anybody or anything we can point the finger at for these discrepancies? How successfully do your models scale up or down in complexity?

Draw diagrams like the one at the top of page 428 for the two cases at the bottom of the diagram at the bottom of the same page (both examples labelled right).