Further Reading: Where Discrete Math Goes Next

This is the map chapter, so its further reading is the most important in the book: it is the set of next books. The chapter (§40.6) already gave the order; this list gives the annotations — what each source is, why it is the right next step, and which chapters of this book it builds on. Do not try to read all of these. Pick the one road that matches your goals and walk it for a year. Tier-1 and Tier-2 sources only.

Step 1 — Consolidate the foundation (now)

Lehman, Leighton & Meyer, Mathematics for Computer Science (MIT 6.042J). Tier 1 — freely available. A second pass over everything in this book, at a complementary angle and with a strong CS flavor. The fastest way to firm up anything that still feels shaky before you go forward. Read the sections matching the chapters you found hardest.

Rosen, Discrete Mathematics and Its Applications (8th ed.). Tier 1. The market-standard reference, with the largest graded exercise bank in print. Use it as a problem source, not a cover-to-cover read — when you want more drill on counting, graphs, or number theory, it is there.

Step 2 — Go deep on algorithms (the highest-leverage sequel)

Cormen, Leiserson, Rivest & Stein (CLRS), Introduction to Algorithms (4th ed.). Tier 1. The natural next book for almost every programmer, and the single highest-leverage step after this one. The recurrences (Chapters 18–19), graph algorithms (Part V), and complexity (Chapters 14, 37) you learned here are its prerequisites, and it is where they pay off in full. Its treatment of linear programming and NP-completeness directly extends §40.1 and §40.5.

Step 3 — Pick one direction and commit

Choose the road that matches your goals. Each entry names the flagship source, what it is, and the chapters it builds on.

Combinatorial optimization and operations research (extends §40.1)

Cook, Cunningham, Pulleyblank & Schrijver, Combinatorial Optimization. Tier 1. A standard graduate text on exactly the genre §40.1 surveyed — LP, duality, matching, flows, and integer programming — by leading authorities. The rigorous home of the "model it as an LP, use duality for bounds" toolkit. Builds on Chapters 32, 34, and §40.1.

Williamson & Shmoys, The Design of Approximation Algorithms. Tier 1 — freely available (authors' PDF). The mature response to Chapter 37's bad news: when exact solutions are NP-hard, prove guarantees on near-optimal ones. LP duality (§40.1) is its central technique. The right book if "NP-hard, but I still need an answer" is your daily reality. Builds on Chapters 37 and §40.1.

Information and coding theory (extends §40.2)

Cover & Thomas, Elements of Information Theory (2nd ed.). Tier 1. The standard graduate text on Shannon's theory — entropy, source coding, channel capacity, and far beyond. The definitive place to turn §40.2's two formulas into a field. Builds on Chapters 26, 38, and §40.2.

MacKay, Information Theory, Inference, and Learning Algorithms. Tier 1 — freely available (author's PDF). A famously readable book that links information theory to machine learning and coding (including the LDPC codes named in §40.2), with vivid examples and exercises. The most enjoyable on-ramp if Cover & Thomas feels austere. Builds on Chapters 26, 38, 20–21, and §40.2.

Ramsey theory and the probabilistic method (extends §40.3)

Alon & Spencer, The Probabilistic Method (4th ed.). Tier 1. The canonical book on the technique §40.3 previewed, by the people who built the modern field. Starts almost exactly where the chapter's Erdős argument ends and goes to the frontier. Builds on Chapters 20–21 and §40.3.

Graham, Rothschild & Spencer, Ramsey Theory (2nd ed.). Tier 2 — a real, standard monograph; exact edition details worth confirming for your copy. The classic dedicated treatment of Ramsey numbers and their explosive growth — the deep version of §40.3's "complete disorder is impossible." Builds on Chapters 17, 27–33, and §40.3.

Programming-language theory and category theory (extends §40.4)

Milewski, Category Theory for Programmers. Tier 1 — freely available (online and PDF). Written for working programmers: it develops categories, functors, and monads (§40.4) using code and types you already use, making it the gentlest serious entry to the subject. Builds on Chapters 9, 24, and §40.4.

Pierce, Types and Programming Languages. Tier 1. The standard text on type systems and the formal theory of programming languages — where the category-theoretic abstractions of §40.4 connect to the proof assistants (Coq, Lean) and type theory the chapter mentioned. Builds on Chapters 4–7, 9, and §40.4.

Theory of computation (extends §40.5, deepening Part VI)

Sipser, Introduction to the Theory of Computation (3rd ed.). Tier 1. The standard, beautifully clear text on automata, computability, and complexity — the rigorous continuation of Chapters 35–37 and the natural deep dive for §40.5's "which problems not to attempt." Builds on Chapters 10, 35–37.

Cryptography (extends §40.5's post-quantum note)

Katz & Lindell, Introduction to Modern Cryptography (3rd ed.). Tier 1. The standard modern treatment: rigorous definitions, proofs of security, and the path from the RSA you implemented in Chapter 25 toward the post-quantum and zero-knowledge frontiers §40.5 named. Builds on Chapters 22–26.

Mathematical depth and technique (sharpens the whole toolkit)

Graham, Knuth & Patashnik, Concrete Mathematics (2nd ed.). Tier 1. Sums, recurrences, generating functions, and number theory developed as working tools with extraordinary care. The book to read if you want your technique — the manipulation behind §40.1's algebra and §40.3's counting — to become fluent. Builds on Chapters 11, 15–19.

A few primary touchstones (optional, for the historically curious)

Shannon, "A Mathematical Theory of Communication" (1948). Tier 1 — freely available; the founding paper of information theory. The original source of §40.2's entropy and capacity. Surprisingly readable, and worth seeing once to appreciate how much of a field one paper can start.

Erdős, "Some remarks on the theory of graphs" (1947). Tier 2 — a real, landmark short paper that launched the probabilistic method; track down a reliable copy. The two-page argument behind §40.3's exponential Ramsey lower bound, in the author's own hand.

Suggested order

  1. Now: skim the Lehman–Leighton–Meyer sections matching your weakest chapters; keep Rosen as a problem bank.
  2. Next 3–6 months: work through CLRS — it is the sequel that makes everything here pay off, and it feeds every Step-3 road.
  3. Then pick exactly one Step-3 road and read its flagship over 6–12 months: optimization (Williamson–Shmoys, free) · information (MacKay, free) · probabilistic method (Alon–Spencer) · category theory (Milewski, free) · theory (Sipser) · crypto (Katz–Lindell).
  4. Optional, anytime: read Shannon 1948 once for the pleasure of seeing a field begin.

One warning, repeated from §40.6: breadth here is a trap; depth compounds. One road, read seriously over a year, will teach you more — and serve you better — than a shallow tour of all six. Use this list to choose, not to sprint.