IdeaWell

Project ideas from Hacker News discussions.

To save C, we must save ABI (2022)

Original Article

Hacker News Discussion

📝 Discussion Summary (Click to expand)

3 Dominant Themes in the Discussion

Theme Core Idea Supporting Quotes
1. C’s “middle‑ground” creates UB and byte‑level expectations Many users point out that C sits between low‑level assembly and high‑level languages, so operations like direct memory access are defined in C but often lead to undefined behavior in practice. > “Not every high‑level language gives you byte‑level access to the representation of memory objects.” — uecker
> “It's not normal in any other language that constant‑folding in the compiler has different behavior than running an expression on the machine.” — titzer
2. ABI stability and evolving standard types The conversation circles around how changing core types (e.g., intmax_t) affects library compatibility and the difficulty of “bottom‑up” ABI evolution. > “Because it imagines that no library other than libc has an ABI that depends on intmax_t.” — aw1621107
> “The mechanism actually does not solve the problem it claims to solve.” — uecker
3. Double‑standard critique of C advocacy Several participants call out a perceived bias: C extensions are praised while similar extensions in other languages are condemned, and straw‑man arguments are common in the debate. > “Because contrary to urban myths, C is a normal high level language like everything else.” — pjmlp
> “Wouldn't be a authentic pjmlp comment unless they shit on C/C++ and/or praise Java/.NET with a bunch of straw‑men :)” — sylware

Summary – The discussion revolves around (1) the inherent UB and byte‑level quirks of C, (2) the practical challenges of maintaining ABI stability when foundational types change, and (3) a heated critique of the double‑standard often shown toward C compared with other languages. These three themes capture the most frequent viewpoints and are backed by direct quotations from the participants.

🚀 Project Ideas

Generating project ideas…

StructMapper

Summary

  • Automatic generation of portable, alignment‑aware byte‑level accessors and packed‑struct layouts from C definitions, eliminating manual offset math and UB risks.
  • Provides safe, cross‑compiler mapping of structs to byte buffers for serialization, network protocols, and low‑level interop.

Details

Key Value
Target Audience Systems programmers, embedded engineers, library authors
Core Feature Input a C struct; output C header + implementation that maps it to a packed byte array with compile‑time alignment checks
Tech Stack Rust‑based CLI + LLVM bindings (optional GCC/Clang plugin)
Difficulty Medium
Monetization Revenue-ready: $19/mo

Notes

  • HN commenters often complain about “byte‑level access to memory objects” (e.g., uecker: “Yes, Rust copied many good ideas from C (and added many new).”) – this tool directly addresses that pain.
  • Enables safe low‑level code without resorting to fragile pointer casts, opening discussion on portable struct serialization.

ABIVersioner

Summary

  • Cloud service that analyzes C library headers, computes ABI diffs, and generates versioned compatibility shims to evolve ABIs safely without breaking downstream code.
  • Offers migration guides and auto‑generated stub libraries for older ABI versions.

Details

Key Value
Target Audience Open‑source library maintainers, firmware vendors, CI/CD pipelines
Core Feature Upload a header; receive ABI version, layout diff, and migration script plus optional legacy stub libs
Tech Stack Python backend, Graphviz diff visualizer, Docker container
Difficulty High
Monetization Revenue-ready: $30/mo subscription

Notes

  • Discussions about “intmax_t isn’t ABI‑stable” and the need for ABI evolution (e.g., aw1621107: “The mechanism actually does not solve the problem it claims to solve.”) highlight the exact gap this service fills.
  • Potential to spark conversation on stabilizing C ABI changes in large ecosystems like Linux kernel modules.

IntrinsicKit

Summary

  • Header‑only, portable library of modern CPU‑intrinsic wrappers (popcnt, clz, bit‑scan‑reverse, etc.) that auto‑dispatches to the appropriate instruction or fallback implementation.
  • Removes the need for inline assembly while keeping high‑performance, architecture‑specific code accessible.

Details

Key Value
Target Audience Game developers, HPC engineers, performance‑critical library authors
Core Feature Generic intrinsic API that selects the best implementation at compile time; extensible for new instructions
Tech Stack C++20 header library, optional Rust fallback, CI testing on x86, ARM, RISC‑V
Difficulty Low
Monetization Hobby

Notes

  • Commenters lament the ubiquity of UB when touching byte‑level data (e.g., titzer: “It’s not normal in any other language that constant‑folding in the compiler has different behavior than running an expression on the machine.”) – IntrinsicKit offers a safe, portable path to leverage advanced instructions.
  • Sparks discussion on making low‑level performance features accessible without sacrificing portability.

Read Later