These are two deeply theoretical books that I expect in these days of watered down first degrees are at second or third degree level. Relatively speaking in my day I would pitch them at the level of third year undergraduate special subject (i.e. electives) choices. You need a sound mathematical foundation to tackle either. This makes it particularly ironic that someone has thoroughly messed up the preface to the Savage book. It states that the book is in four parts, goes on to describe a book in three parts and though it has the chapter titles correct it attributes them to the wrong parts of the book. This lack of attention makes me wonder just how much care was taken over the rest of the book. While my maths and computing theory is up to reading books such as these, it is not up to the level where I can validate technical detail. I have to trust the author. The state of the preface leaves me feeling insecure.

Those responsible for the design of the machines we use should be familiar with the material covered. I bet many are not. The detailed consideration of trade-offs between space and speed are revealing. The entire theoretical framework exists to tackle issues of caching, multi- level caching etc. all the way down to use of secondary storage. Of course studying this sort of material is hard work and way beyond what anyone could expect of the working programmer. However knowing that the material exists allows us to be more critical of the inadequacies of many designs. We may not have the time to determine the answers for ourselves but we no longer need to accept the excuse that no-one knows how to do that.

I wish I could justify a greater depth of discussion of the relative merits of these two books but neither space nor the interests of most readers can justify that. If I had to choose between the two books my vote would go to Gregory Taylor's. His English is just that much more readable. However either would be a valuable study for those aiming at mastery of Computing Science.