Devlog
This is a public research journal. Not a manifesto. Not a pitch deck. A journal.
How we work
We use the scientific method. Not because it sounds impressive — because it is the only honest way to build something new.
We observe. We look at how software systems describe themselves today. RPC, REST, GraphQL, gRPC, MCP, Smithy, CloudEvents, AsyncAPI. We read specifications. We study history. We notice patterns.
We hypothesize. When the same structure appears independently in seven different systems — we hypothesize that the structure is fundamental, not invented. When every operation in every system has input, output, and the possibility of failure — we hypothesize that this is physics, not convention.
We test. We build. A schema. Type definitions. A playground. If the hypothesis survives contact with code — it stands. If it breaks — we record why and move on.
We record. Every decision, every dead end, every surprise — written down in the order it happened. Devlogs are append-only. We do not edit old entries to look smarter. Entry #4 may contradict entry #18. That is not a bug. That is research.
What we claim
If we cannot find a counterexample to a statement, we record it as a fact. Not as truth — as a fact that has survived every test we could throw at it. If you find a counterexample, open an issue. We will update the journal. That is how science works.
What we do not claim
We do not claim to be the first. Berners-Lee, Fielding, Sadalage, Kleppmann, Vint Cerf, Bob Kahn — they built the foundations. We walk on their roads. We are not architects. We are explorers who noticed that several roads lead to the same place and asked: what is at that place?
We do not claim to be right about everything. Some entries in this journal will age poorly. Some hypotheses will be disproven. That is fine. A journal that is never wrong is not a journal. It is marketing.
Who we are
Researchers. Dreamers. Builders. Working in public, with zero stars and zero funding. Asking one question: what if the operation is the fundamental primitive of computation, and nobody wrote it down?
Maybe the answer is yes. Maybe the answer is no. The journal will tell.
Want to challenge an entry or write the next one? Open a PR.