Files
Aura/fieldtests/cycle-300-document-first/df_3_bare_plateau.md
T
claude 63711a1d3a fieldtest: document-first — 4 examples, 7 findings
Downstream-author fieldtest of the #300 surface, public interface only,
real GER40 archive. All four axes hold: the closed
generate -> show -> hand-extend -> re-register -> run loop (2-cell
extended campaign), the typed stop grammar's clap rejections, bare
plateau's content-id identity with plateau:mean, and the show read-back
(byte-identical, no framing) with a clean unknown-id refusal.

Findings routed: 1 bug (show outside a project blames the id instead of
the missing Aura.toml — RED-first fix follows), 2 friction items filed
to the tracker (opaque process-document select refusal; no prefix-id
hint on show), 4 working (carry-on). The fieldtest spec survives
git-ignored as the next planning cycle's reference input.

refs #300
2026-07-21 13:16:59 +02:00

2.2 KiB

df — example 3 — bare plateau vs explicit spellings (Axis 3)

Task: confirm the CLI --select plateau sugar is content-id-identical to --select plateau:mean, distinct from the other spellings, refuses garbage cleanly, and check whether the alias carries into a hand-authored process document. Binary @ a12617bb.

CLI --select: content-id identity (walkforward over GER40, same --name)

campaign8 = first 8 hex of the generated campaign content id:

--select plateau        -> campaign8 = e35e493e
--select plateau:mean   -> campaign8 = e35e493e   <- IDENTICAL
--select argmax         -> campaign8 = eeede651
--select plateau:worst  -> campaign8 = ea1f4e3a

campaign runs confirms plateau and plateau:mean collapse to ONE stored id (e35e493ea4a9...d13acc, run 0 + run 1), argmax and plateau:worst are separate ids. Confirms the cycle claim: the two spellings generate a content-id-identical campaign.

CLI refuses an unknown spelling (usage lists the accepted set)

$ aura walkforward ... --select plateau:bogus
aura: Usage: aura walkforward ... [--select <argmax|plateau|plateau:mean|
      plateau:worst>] ...
$ aura walkforward ... --select garbage        (same usage refusal)

The --select help string also documents it: argmax | plateau (alias for plateau:mean) | plateau:mean | plateau:worst.

the alias does NOT carry into a process document (df_3_..._REFUSED.json)

$ aura process validate df_3_select_plateau_process_REFUSED.json   [exit 1]
aura: process document: block std::sweep: unknown select rule "plateau"

$ aura process validate <same doc with "select":"plateau:mean">    [exit 0]
process document valid (intrinsic): 1 pipeline blocks, 0 gate predicates

Outcome: the CLI-flag behaviour and identity claim are WORKING and documented (help + usage). The FRICTION is the document-side refusal: unknown select rule "plateau" neither enumerates the accepted rules (argmax | plateau:mean | plateau:worst — as aura process introspect --block std::sweep would) nor bridges bare plateau -> plateau:mean. In the document-first hand-extend workflow this cycle is about, an author who learned bare plateau at the CLI will paste it into a document and hit an opaque refusal.