out/RETURN-C5.md
Saved at C5.C5d-p2: table, model and figure. This reading copy substitutes participant identifiers in saved text; it is not a new interview response. Reading Markdown.
Participant handles in these reading copies use the pixel stream's identifiers: base64(SHA1(username)), case preserved, without a salt. These are linkable pseudonyms. Profile links are removed; missing-author markers remain. Original research files are unchanged. Quote offsets refer to original text. Source and reading-copy hashes are listed separately.
Source commit: 08d559e79c5b, 2026-10-15T17:01:12+02:00. Saved-state date: 2026-10-15T15:00:34Z. Catalogue close receipt Commit date and saved-state date are distinct.
Historical account, 15 October 2026. D3 revised R01 and P05, bounded A21 and added fifteen candidates. The 24 October D3c head says the C5 account has not yet been rewritten for those amendments. Read the D3 handback alongside this earlier account. D3 handback, 24 October.
The run later retold the model and audited its limits. Read the final D6 account and E1/E2 audits for the completed run. Completed v43 results.
Original GitHub file. Original SHA-256: 2a659718e7d61bc6bf8d2287953aaa67752bffa2e943ceb1dc750aefff2853fc. Reading-copy SHA-256: 2a659718e7d61bc6bf8d2287953aaa67752bffa2e943ceb1dc750aefff2853fc.
RETURN-C5: the return specification for stage C5's three readers
Frozen with the stage file. The dispatch serves it beside each of the
packets C5-reader-1, C5-reader-2 and
C5-reader-3, and the validator's schema is read from it.
One reader, one packet, one return.
What you return
One JSON line, then the end marker on a line of its own. Nothing else, before or after.
reader: your reader id as given to you.served: the sha256 ofaccount.md, copied exactly from the packet's first page (64 lowercase hex characters).claims: five to nine claims you takeaccount.mdto make, each in your own words, one sentence each.relations: the relations between elements you take the account to claim, each as a short phrase with the ids, such as"A03 acts on P04"or"P05 comes before P08". An empty list if you take it to claim none.could_not_follow: each sentence you could not follow, and each term a reader would not accept as plain words, quoted exactly from the packet. An empty list if there is none.unsupported: in a few sentences, the claims of the account you cannot find the evidence for in the packet, and any room you can name that the account does not fit. An empty string if there is none.rooms: one object for each of the three rooms you were given, each with:sub: the room's name as the packet gives it, withoutr/;fits:trueif the account describes what happened there,falseif it does not;instead: one paragraph. Where the account fits, say which of its parts you saw in the room. Where it does not, say what the room did instead.
declarations: one object for each sentence served to you under "The sentences in question", each with:id: the declaration's id as the packet gives it (T4-NNNorDNN);sentence: the sentence, copied exactly;stands:trueif the sentence stands under the rule you were given,falseif it does not;why: one sentence: what supports it (a count, a finding id, a setting fact), or which part of the rule it fails.
codes: one object for each record served to you under "The records to code", each with:record: the record's citation, copied exactly (sub/cNNNN);code: exactly one ofM,MP,PorO, by the code definitions you were given.
{"reader": "...", "served": "<sha256 of account.md>", "claims": ["..."], "relations": ["A03 acts on P04"], "could_not_follow": ["quoted"], "unsupported": "...", "rooms": [{"sub": "...", "fits": true, "instead": "..."}], "declarations": [{"id": "T4-005", "sentence": "...", "stands": true, "why": "..."}], "codes": [{"record": "sub/c0001", "code": "M"}]}
END-OF-RETURNA disagreement is never a verdict: your reading decides nothing.
After your line, write the end marker END-OF-RETURN on a
line of its own. Confirm that you reached the end of the packet by
writing that marker; a return without it is a partial return and is
reissued.