Is there an equivalent to Feynman's lectures on physics but for computer science?
Which books about computer science will remain relevant in one hundred years, when most engineers will probably be using programming languages and frameworks whose creators are not even born yet? I don't work in software, but I like to collect foundational texts in various disciplines. I have some programming experience under my belt, and I've taken a few introductory classes about data structures and algorithms, but my personal library lacks CS texts.
The obvious answer is Structure and Interpretation of Computer Programs by Abelson and Sussman, often called SICP. It is often discussed on HN and is generally regarded as a classic with great staying power. It appeared in 1985 so it already has a good start on a hundred years.
It has some features in common with The Feynman Lectures: It presents a very personal, not always mainstream, view of its subject. It was originally written as a first-year textbook (for MIT instead of Caltech) but many in its orginal intended audience found it too difficult or just too weird, and it is most appreciated now by experienced programmers and computer scientists.
I can understand calling these "texts", but "books"? These remind me more of the sectioned documentation pages that one often comes across on the "document web"; e.g. https://tldp.org/HOWTO/Plug-and-Play-HOWTO.html
To me, "book" implies constant-sized pages and content formatted such that it could be used to produce an actual hardcopy book (e.g. a PDF).
What a great thing. I've been working on an interactive textbook for React and TypeScript, mostly just to better support one of my courses. I feel like faculty should be encouraged and supported to produce free, open materials like this.
The three books I clicked through several chapters of were unpleasantly focused on the features of the C# standard library and details of network administration in Digital Ocean (?!?).
Tim Roughgarten of Stanford has split his big algorithms book into four $20 paperbacks, so you can pick and choose an area you'd like to reinforce your knowledge of for not too much money.
Is there a resource that delves heavy into the application of the OSI model? Specifically, on how tcp actually works under the hood. Or how applications impliment their netcode ontop of udp? Typically we'd just say recv() and get data, but that doesn't detail the underlying process of how packets are handled.
Things don't follow the OSI model nor have they ever really, it never caught on. The 4 layer TCP model is much more realistic (and less formalized) but even then there can be differences sometimes. You can spend a lifetime learning OSI to only end up finding a few real world protocols like IS-IS in use and co-opted to be used outside a full OSI stack anyways.
Regardless knowing the layering model to the letter tells you very little about how TCP works for instance. After all that's the reason for layers, you don't need to know how the protocol does it's thing just where it fits in the stack and what happens between the boundaries (which in the TCP model is "not much"). The other commenters suggestion of Stevens book has been the traditional recommendation though I will say it hasn't sat idle since then and you !ay want to follow it up with some more recent resources to get a complete picture of modern TCP extras.
For the UDP one also look into the design of QUIC for ideas in that space.
I love to see this kind of stuff, especially when the books are used in the curriculum. Textbooks are a racket and K-State, like so many state schools, struggles to be affordable to the student of the state. This kind of thing can help a lot.
Is there an equivalent to Feynman's lectures on physics but for computer science?
Which books about computer science will remain relevant in one hundred years, when most engineers will probably be using programming languages and frameworks whose creators are not even born yet? I don't work in software, but I like to collect foundational texts in various disciplines. I have some programming experience under my belt, and I've taken a few introductory classes about data structures and algorithms, but my personal library lacks CS texts.