What This Book Actually Covers
Information and Coding Theory by Jones is a university-level textbook that sits somewhere between an introduction to Shannon's work and a practical guide to constructing error-correcting codes. It's aimed at undergraduates who already have some background in linear algebra and discrete mathematics, though not everyone who picks it up realizes that upfront. The chapters move from the basics of entropy and channel capacity into Huffman coding, then shift into block codes, cyclic codes, and Reed-Solomon before touching on convolutional codes and Viterbi decoding. That last section is where the book gets genuinely useful for anyone who has ever tried to understand why their satellite link dropped packets during a solar flare. The reason this particular text stays on syllabi is that it doesn't treat coding theory as pure abstraction. Jones works through proofs but then ties each concept back to implementation concerns. Most other textbooks either drown you in measure-theoretic probability or skip straight to MATLAB without showing the math underneath. This one sits in the middle, which is both its strength and its weakness. I've used this book as a reference when designing parity-check schemes for embedded systems. The walkthrough on syndrome decoding for Hamming codes is solid. The section on generator polynomials for cyclic codes is equally reliable. But I've also run into gaps that frustrated me enough to pull a second reference off the shelf. For example, the treatment of finite fields is terse to the point of being almost hostile. If you've never seen GF(2^m) construction before, you will stall at chapter four. I spent about two days cross-referencing with Lin and Costello before the Galois field material clicked. That's a pretty common pattern with this book.
How to Actually Use This Book
Start by skimming the first three chapters before committing to the full read. You need to know whether the notation matches what your program already uses. Jones writes checksums and parity checks in a way that assumes you're comfortable with matrix notation over binary fields. If that's not your default frame of reference, you'll waste time translating between his notation and whatever you're working with. The worked examples are where the book earns its keep. Don't skip them. Each code construction comes with a concrete example, usually using small parameters like a (7, 4) Hamming code or a (15, 11) cyclic code. These are intentionally small so you can trace the encoding and decoding by hand. I find that tracing a single codeword through the full encoder-decoder cycle manually is the fastest way to internalize the mechanics. It takes about ten minutes per example but sticks with you longer than any amount of passive reading. When you hit the Reed-Solomon section, expect to slow down significantly. That material assumes you're comfortable with polynomial arithmetic over extension fields, and Jones doesn't hold your hand through that transition. I recommend having a supplementary source on finite field arithmetic open alongside. Berlekamp's algorithm for error locator polynomials is covered, but the derivation is compressed. If you're implementing this yourself rather than just studying it, you'll want to derive it yourself too.
A Problem I Ran Into and How I Fixed It
While working through a project involving CRC-based integrity checks, I noticed that the textbook's treatment of cyclic redundancy checks didn't address a specific edge case I hit: what happens when your message length is shorter than the generator polynomial degree. Jones mentions this briefly but doesn't walk through the padding behavior that real protocols use. In practice, you need to left-pad the message polynomial with zeros so it reaches the same degree as the generator before division. Without that step, the remainder operation produces an incorrect CRC value. I spent about an hour debugging a test case where the CRC output looked perfect for long messages but failed consistently on short ones. The fix was simply ensuring the input message was zero-padded to length n before running the division. Once I added that step, the implementation matched the theoretical remainder exactly. This isn't something the book emphasizes enough, and it's the kind of detail that shows up in real implementations all the time.
Get the Full Details
Where the Book Falls Short
The convolutional code section is the weakest part. It covers the trellis diagram and Viterbi algorithm but doesn't go deep into soft-decision decoding or iterative decoding variants. If your work involves anything beyond basic hard-decision Viterbi, you'll need to look elsewhere. Turbo codes and LDPC codes get a mention at most. That's a significant gap if you're working with modern wireless standards or storage systems that rely on LDPC for error correction. The sections are also uneven. Some chapters have well-designed problems that reinforce the material. Others feel like they were assembled from old exam questions without much editorial polishing. I'd recommend skipping exercises that don't clearly connect to the preceding text and focusing on the ones that ask you to construct or analyze specific codes by hand.
Who Should Read This
This book works well as a primary text for a senior undergraduate course or as a supplementary reference for graduate students who need a refresher on classical coding theory. It's less suitable as a standalone introduction if you've never encountered information theory before. The prerequisite knowledge isn't spelled out clearly in the preface, so readers sometimes discover too late that they need stronger foundations in probability and linear algebra. If you're looking for a book that covers modern coding theory including turbo codes, LDPC, and polar codes, you'll want something more recent. For that, refer to Richardson and Urbanke's work or the more current editions of Proakis. But for a solid grounding in the classical material that underpins everything built on top of it, Jones remains a dependable choice. The prose is clear, the examples are practical, and the mathematical rigour is appropriate for the intended audience.
Getting a Copy
The book is available through most academic publishers and major retailers. University libraries typically carry it, which is worth checking first since the price can be steep for students. Digital versions exist through academic platforms, though the formatting in some ebook editions makes the mathematical notation harder to read than the print version. If you plan to work through the exercises with pencil and paper, the physical copy is the better experience.
