csapp-zh-markdown: a Chinese CSAPP edition with the labs and the answers

One directory per chapter, twelve chapters in total

csapp-zh-markdown is a Chinese language study edition of Computer Systems: A Programper's Perspective, the third edition, laid out as plain Markdown in a repository anyone can clone. The structure is one directory per chapter, from the first chapter on a tour of computer systems through the chapters on information representation, machine level programs, processor architecture, performance optimisation, the memory hierarchy, linking, exceptional control flow, virtual memory, system level I/O, network programming and concurrent programming. The preface, the error handling appendix and the full bibliography sit alongside them. The three parts of the book keep their own introductions at the head of chapters 2, 7 and 10, so the part structure survives the split into files. The project states that the text is kept as faithful as possible to the original wording and order, that ordinary text, code, tables and simple formulas are editable text, and that diagrams needing spatial structure are kept as local images.

Two ways into every chapter: chapter.md and README.md

The most useful design decision is that every chapter directory offers two entry points, and the project explains the difference precisely. The `README.md` in a chapter directory is only a navigation page, a table of contents for the subsections, while `chapter.md` holds the continuous text of the whole chapter. Reading front to back means opening `chapter.md`, and looking something up means opening the `README.md` and picking a subsection. The subsection Markdown files and their images are kept as well, so nothing is lost in either mode. The project walks through the intended flow with a concrete example: when studying two's complement, you click into subsection 2.2.3 from the chapter index, work the exercises as you read, and then check your answers through the answer entries at the head and foot of the page. Notes are taken by editing the local Markdown files or keeping a separate notes file, which is the reason the files are distributed as plain text rather than as a single document.

238 answers, kept apart from homework and lab work

The exercise answers are the original answers printed in the book, and the project is careful about their scope. The table of contents states that there are 238 of them, and that each chapter has an independent exercise answers entry at the head and the foot of the page so you can check your work after finishing. The answers page also states plainly that these are the original exercise answers provided by the book, and that the project does not include self written homework solutions or lab solutions. That boundary matters for anyone using this repository, because the exercises inside the text include the homework problems, and those stay blank on purpose, with the fill in the blank tables keeping their empty cells so a reader can work them in the file. Where the book contains typographical problems, they are recorded separately in an answer proofreading note rather than silently corrected in the answer text, so a reader can see what was verified and what was left as printed.

Eight labs with Chinese guides and the official tar packages

The lab half of the repository is a translation of the eight self study labs, each with a Chinese document and the original package next to it. The labs are Data Lab for bitwise and floating point work, Bomb Lab for assembly reading and debugging, Attack Lab for stack buffer overflows and return oriented programming, Architecture Lab for Y86-64 and pipelining, Cache Lab for cache simulation and matrix transposition, Shell Lab for processes and signals, Malloc Lab for dynamic memory allocation, and Proxy Lab for networking and concurrency. Each lab directory contains the official self study `.tar` package, and the project also ships a Chinese guide to the Y86-64 processor simulator used in the Architecture Lab. The documented procedure is three steps: read the Chinese description of the lab, take the `.tar` file from the link at the top of the document and unpack it into your own working directory, then read the differences between the self study package and the original instructions, and configure, build and test according to the lab instructions and the guide inside the package.

What the compatibility and provenance pages admit about the packages

Two documents in the lab directory set expectations honestly, and they are the first files to read. The compatibility page lists where the self study packages differ from the original course documents. The project states that the packages mainly target a Linux environment, that some use older toolchains or interpreters, and that course dates, server addresses and submission paths retained in the text come from the official template and are not services this repository provides. One specific gap is called out: the Malloc self study package includes only two short tests and does not include the complete grading tracking file. The provenance page records the official source of each package and its SHA-256 checksum, so the download can be verified rather than trusted. For a self study setup this is the difference between a package that is known to differ from the course and one that is quietly outdated, and it is also why the project keeps the description of each difference in a file rather than in a paragraph buried in a readme.

A reading site that shares one body of text with the repository

Version v1.2 added an online reading site, and the project is explicit that the site and the repository share the same text. The site is described as offering chapter navigation, full text search, previous and next page jumps and a dark mode, working on desktop and phone, and requiring no download and no software. The repository remains the primary artefact and is meant to be kept alongside, so the Markdown files can be read offline and annotated, and you get the whole set with a clone.

git clone https://github.com/SunnyMaria/csapp-zh-markdown.git

After the clone, you open the root `README.md` in any editor that previews Markdown. The project says it will keep improving text verification, image quality, the search function and the reading experience on computers and phones, and that the Markdown files will continue to be retained and shared with the site. Releases are numbered v1.0 on 2026-09-16 for the Chinese Markdown study files, v1.1 on 2026-09-20 for reading structure improvements and a full proofreading pass, and v1.2 on 2026-09-21 for the online site, with the last push on 2026-10-09.

Where the text came from, and the licence gap

The provenance of the content is stated in the acknowledgements and it has consequences. The project credits the Hansimov/csapp repository, and says the OCR source files used for the text come from there, providing the basis for the text checking, the chapter splitting and the Markdown conversion. The authors of the original book and Pearson are credited separately, along with the official book site for the preface, figures, labs and other study files. What is not stated anywhere in the project documentation is a licence. The top level of the repository contains no licence file, and the project makes no licensing claim for the translated text, the images or the answers, even though the content is a translation of a copyrighted book and the official site is named as a source for the figures and labs. For a study repository used privately this may not matter to anyone, but for anyone who intends to redistribute the files, host it, or build on it, the absence of a licence means there is no permission to rely on, and the question has to be settled with the rights holders rather than with the repository.

When this repository is the wrong place to learn from

Three mismatches are worth naming. The first is language. This is a Chinese edition, so a reader working in English gains nothing from it, and the project says nothing about other languages beyond noting that multi language collections will be evaluated as the work progresses. The second is the answers. The repository supplies the book's own exercise answers and nothing beyond them, so a reader who wants worked lab solutions, which is a large part of what the self study packages normally assume, will not find them here. The third is the tooling. The labs are compiled and tested on a Linux environment with older toolchains, the course dates and submission paths in the text point at infrastructure this project does not run, and the Malloc package is missing its full test set. A reader with a course timetable, a grading script and a current toolchain is better served by the course documents, and this repository is at its best as a reading, lookup and translation aid alongside them.

Editorial conclusion

Adopt csapp-zh-markdown if you read CSAPP in Chinese and want a single searchable repository that carries the text, the original exercise answers and the lab documentation in one place, with the official self study packages kept alongside the translations. Skip it if you need the book in a language other than Chinese, if you want worked lab solutions, because the project states it deliberately does not include those, or if you need a licence you can rely on, since the repository carries no licence file and the content is a translation of a Pearson publication. Verify three things first: that the chapter directory you open contains the `chapter.md` you expect rather than the navigation `README.md`, that the lab package you download matches the SHA-256 values listed in the packages provenance page, and that your Malloc lab expectations are adjusted, because that package ships only two short tests without the full grading tracking file.

Frequently asked questions

Official sources

1. Issues
2. README
3. Releases
4. SunnyMaria/csapp-zh-markdown on GitHub