Dimensional analysis — This kind of works directly for CS. At least the idea that most estimation problems can be expressed as a single product of terms.

“Taking out the big part” & “Easy cases” — These can also be phrased as “ignore the corner cases”

Lumping — Hard code things, Look-up tables,

And given that we have computers available, things like successive approximations (light weight GA and optimization techniques) and time step simulations (e.g. for queuing theory) should be covered. Also, I think some of the same mindset can be used not just for estimation (a symbolic answer to “can this work”) but for experimental prototyping (an empirical answer).

The street fighting ideas might give students a basis to conceptualize and think about the solutions to problems before sitting down and typing up a prototype.

Although you may know why these tidbits are extremely useful to have readily accessible, I’m not sure they will stick without the context of experience. Students will normally not have enough experience, whereas someone seeking out and reading a book will.

As for content: Definitely need to include the Pareto Principle. You might also want a section on social issues related to CS, such as “The Mythical Man-Month” and “Pournelle’s Iron Law of Bureaucracy.”