README.md

coccoc

What This Is

coccoc is a practitioner's changelog. Not a tutorial site. Not a blog. A running technical log for people who are actually building things and have opinions about the tools they use.

The structure reflects how developers actually think about information:

  • Dispatches — Opinionated deep-dives. The kind of thing you'd write after spending a week with a tool and having actual opinions about it. Not how-tos.
  • Stack Files — Structured breakdowns of real-world developer setups. The exact stack a solo dev used to ship something, documented honestly.
  • Terminal Log — Short raw notes. Commands, discoveries, one-liners that saved hours. Like a public TIL but with more context.
  • Benchmark Lab — Honest comparisons. Real numbers, stated methodology, no vendor copy. Each benchmark is standalone.
  • OSS Radar — A curated watchlist of repos worth following, with brief honest takes.

Contact: [email protected]

Who Writes This

M
@mschubert_dev
Principal engineer. Does too much infra for someone who claims to prefer application code. Has strong opinions about observability tooling that he expresses at length in Dispatches and Stack Files.
R
@rkovacs_ml
Former academic, now AI tooling obsessive. Spent three years in NLP research before deciding she'd rather evaluate models than train them. Responsible for Benchmark Lab and most of the OSS Radar.

Editorial Rules

These aren't aspirational. They're constraints we actually hold each other to:

  • No "game-changing". No "revolutionary". No "seamlessly".
  • State when something is an opinion vs a measured result. Be explicit.
  • If you ran a benchmark, describe the methodology before the numbers. Always.
  • "We tried X for three months" is a sentence. "Many developers find" is not.
  • We only review tools we actually use. No unvetted vendor copy.
  • Benchmarks are labelled "internal testing". Run your own and tell us if we're wrong.
  • Dates are real. If something was written in January and updated in April, both dates appear.

How To Contribute

We're not running an open submission queue. But if you've done something genuinely interesting — built a real setup, run a real comparison, found a command that actually matters — reach out at [email protected].

What we're looking for: specificity. "I compared X and Y on my M3 Pro with this exact workload" beats "X is generally better than Y" every time. We'll edit for clarity but not opinion.

What we're not looking for: tutorials, beginner content, product announcements, anything that reads like vendor documentation.