Harmonize CS core team knowledge & culture

[yvan-sraka]

Originally posted by @yvan-sraka on internal Discord:

I was thinking about « how to improve the overall PR review quality » and how to harmonize our codebase expectations (and find consensus in the writing of the CONTRIBUTING.md), by fixing CS wrong believes in the team.

I can start sharing few resources I found really helpful on my on personal way to a better software engineer mindset:

Books

• « Rust for Rustaceans » is the most up-to-date, accurate and idiomatic resource to level up our language knowledge (the book also tackle a lot of non-Rust specific topics, I could only encourage everyone to read the Chapter 8: Asynchronous Programming and Chapter 10: Concurrency and Parallelism). What about ordering a dead-tree version on no starch press https://nostarch.com/rust-rustaceans to let the book available in Deskopolitan office?!

• « Algorithms » by Jeff Erickson http://jeffe.cs.illinois.edu/teaching/algorithms, because from all the algorithms' university books I read (remind that I spend too much time doing ranked contests, train students to these contests, and write terrible contests subjects) that is the one with the best focus on practical « problem-solving », as say as « finding new original algorithmic solution to new problems combining most common technics of reasoning »

Twitter & Youtube (Official Rust Discourse, Discord and Reddit communities are good places to join too)

Side to the « This Week in Rust » https://this-week-in-rust.org link must read newsletter, I want to advise you to follow some core Rust contributors:

• David Tolnay https://twitter.com/davidtolnay, that wrote most of the beloved Rust crates, like serde, semver, anyhow/thiserror or syn/quote/proc-macro2 and challenge your Rust edgy knowledge with https://github.com/dtolnay/rust-quiz and https://github.com/dtolnay/case-studies

• Steve Klabnik https://twitter.com/steveklabnik and Carol Nichols https://twitter.com/carols10cents that wrote The Book https://doc.rust-lang.org/book/

• Mara Bos https://twitter.com/m_ou_se, that lead the library team and give really good overview on new language feature that land in stable every 6 weeks, also available in https://blog.rust-lang.org/inside-rust/

• Jane Lusby https://twitter.com/yaahc_ that give a really highlighting talk on error management in RustConf 2021 https://www.youtube.com/watch?v=rAF8mLI0naQ&list=PL85XCvVPmGQijqvMcMBfYAwExx1eBu1Ei

• Lily Mara https://twitter.com/thelily_mara "The Importance of Not Over-Optimizing in Rust" RustConf 2022 https://www.youtube.com/watch?v=CV5CjUlcqsw&list=PL85XCvVPmGQgACNMZlhlRZ4zlKZG_iWH5

• You don't know what's OIBITs, Auto traits, HRTBs, GATs, ZSTs, DSTs means ~> https://www.youtube.com/watch?v=6Qi5-VU-kS0 from https://twitter.com/RustWroclaw is a bit outdated but still fun to watch :slight_smile:

• Felix Klock https://twitter.com/pnkfelix (that teach me Rust in 2014 in Paris Mozilla Space) explain Subtyping in Rust https://www.youtube.com/watch?v=fI4RG_uq-WU, that is the core of Rust standard types ergonomics and this video remove the magic you feel for e.g. every time you try to borrow a String

• Last one, for the ones that likes Static Analysis / Formal Verification / Language and Type Theory formula porn https://www.youtube.com/watch?v=iAs0gZ8o0oQ&list=PL85XCvVPmGQiF26DB84AK3i1yjKu52fq_ that highlight the incredible advancements of Rust Belt https://plv.mpi-sws.org/rustbelt/ WG

All that is maybe a bit Rust-focused 🦀 but should help both newcomers and historic members 🦕 of the team, feel free to augment this list with your own useful resources, I can easily turn it into a GitHub discussion ✨

I could continue with some more resources (less Rust focused) about performance and systems:

Articles

• « What Every Programmer Should Know About Memory » https://people.freebsd.org/~lstewart/articles/cpumemory.pdf that really gave (with too many details) insights about how wrong programming pattern against memory locality is a bottleneck of CPU performance (before even thinking about SIMD or other optimizations), understanding a bit of processor cache invalidation mechanisms!

• I think it also always worth the detour to read and understand this wonderful SO top post https://stackoverflow.com/questions/11227809/why-is-processing-a-sorted-array-faster-than-processing-an-unsorted-array about branch prediction fail :slight_smile:

• An article about the not well known layout biases measurement https://users.cs.northwestern.edu/~robby/courses/322-2013-spring/mytkowicz-wrong-data.pdf that teach us that performance benchmarks is a problem that live on top of reproducibility (random NixOS ad here https://r13y.com)

Misc

• Pro Git https://git-scm.com/book, is still THE only proper way to construct a valid mental model on how git actually works behind the hood

• https://12factor.net/ principles, which is must-read guidelines 📑 @damip