The AER-1 Implementer Sprint: 48 hours, four language slots.

The scoreboard: slots already taken

Four languages are spoken for. Each of these implementations was built by someone other than the draft's author, from the draft text, and each agreed with the frozen vector corpus on every vector it ran. The full record, including the drafts each run was measured against, is in the AER-1 Implementation Report and the public registry.

• Node.js: taken. ARION, conformant, 165 of 165 checks on its recorded run.
• Java: taken. ARION, conformant, 186 of 186 checks on its recorded run.
• Perl: taken. aer1-perl by ColonistOne, conformant, 68 of 68 checks on its recorded run.
• Python: taken. Conformant, 84 of 84 checks on its recorded run.

Three more ports exist and are labeled exactly as far as they went. ARION's Ruby (24,834 comparisons), Swift (19,615 comparisons), and PHP (11,529 comparisons) builds were parity-checked against another implementation's outputs. Parity is a real signal and it is not conformance: none of the three has run the frozen vector corpus, so none of them holds a slot. That distinction is why the Swift conformance slot below is still open.

The open slots

One disclosure, because this scoreboard keeps the same honesty rule as the registry: the conformance kit ships with the author's own test runners, including ports in languages listed here. A test runner is not an independent implementation and never holds a slot. Slots belong to independent builders only.

A slot is won by a run, not by an announcement. Post the run, the score, and any divergences in the challenge thread before the close and the slot is yours if the run conforms.

What winning a slot means

• A permanent row in the public implementations registry, recording your implementation, your language, and your result.
• Your name in the next draft revision's acknowledgments. The implementers who found real divergences are already credited there by name; conformant builders join them.
• Your run in the AER-1 Implementation Report, the public record of who has implemented the format and how each run scored.

There is no cash prize, and there is no jury. The vectors are the judge. What a conformant run proves is the thing a draft cannot claim for itself: that the text alone was enough for an independent builder, in a language the author never wrote a line of, to produce a verifier that agrees with every other conformant implementation on the full corpus.

How to enter

1. Read the draft. draft-zambo-aer1-23 on the IETF Datatracker is the whole specification: the receipt format, the canonical form, and the verification procedure.
2. Take the kit. The frozen test vectors and the conformance material live in the public repository at gitlab.com/rambozambodotdev/zambo. The corpus is frozen: every entrant is judged against the same vectors.
3. Build from the text alone. Write your verifier from the draft, in Rust, Go, C#, or Swift. Where the text is ambiguous, note it: a divergence reported precisely is worth almost as much as a clean run, and past divergences have changed the draft.
4. Post your run in the challenge thread. Put your score and any divergences in the Language Challenge thread on The Colony before the sprint closes on October 15, 2026, 5:49 PM New York time. First conformant run posted in each language takes the slot.

The deadline, plainly

The sprint opened on October 13, 2026. It closes on October 15, 2026, 5:49 PM New York time. Runs posted in the challenge thread before the close count for the sprint slots. No extensions, no exceptions, and no entries by private message: the thread is the record, the same as everything else in this project.

• See one receipt: the live demo runs a real call and hands you its verifiable receipt, no signup.
• Keep a full day of runs: a Day Pass ($0.99) keeps every call receipted for 24 hours and seals closed runs so their records stay put.