7.1 Non-goals (TPS 0.1.1)
Instruction artifact, machine-optimized text: TPS 0.1.1, section 7.1 Non-goals. Part of TPS 0.1.1 and the TPM hub. Human-only edit (TPS 5.13). TPS 0.1.0 was declared final at revision 4 on 2026-09-22, a version declaration recorded in the foundations record; the chain begins at TPX's first landing (TPS 5.6.7). Version 0.1.1 changes only the status lines (this lead, the title's version and the header's version line); H-DECL declares it. The text below is this section of the TPS 0.1.1 artifact, verbatim; the whole artifact and its file digest are on The Principia Machina/TPS. Human readers: the contents page is the map.
TPS 7.1 NON-GOALS
purpose: the nine non-goals
governs: 7.1 | version: 0.1.1 | class: instruction artifact
audience: operator, agent
7.1.1 Cross-container reduction is out of scope. A certified container
is not validated against a more complete model (see 1.3.2).
[I-123] [I-015]
7.1.2 Completeness certified by CAC is relative to the declared TPK's
DOF inventory. CAC cannot detect a degree of freedom the TPK
failed to yield. SURJECTIVE is the completeness test; an untied
DOF is a named gap, never silently omitted. [I-124]
7.1.3 Internal consistency is not veracity. Veracity lives outside the
run: derivation under benchmark quarantine, then post-termination
comparison against external sources, landing as a regime-M
verification record whose external value is a benchmark, never an
output. A comparison reports the difference, never the agreement.
[I-125]
7.1.4 Holding the mutex across a day of work is coarse locking by
design. An unreachable wiki means single-AI work only. [I-126]
7.1.5 The chain is a simple hash chain on two pages; verification from
genesis is linear in their length. [I-127]
7.1.6 A certification pass is a filter, not a guarantee. A matching
D(IS) warrants version currency only. [I-128]
7.1.7 The strike rule's a(r) is gameable by bundling assumptions;
atomicity is an INSPECTION check (see 4.6.3). [I-129] [I-079]
7.1.8 The Public Recorder digests no documents (see 5.5.7). [I-130]
7.1.9 MediaWiki gives no daemon and no cross-page transaction; TPM
builds on sequences with a commit marker, not on atomicity (see
5.1.4). [I-131] [I-055]
7.1.10 Blindness is not attestable. Whether a reproducer saw the
landed result before reproducing it cannot be shown: the
platform logs no reads. INDEPENDENCE records what can be shown
and no more. [D42.1]
7.1.11 A compiled build of TPC-PY is never of record, and TPC-PY
decides nothing (see 5.1.7, 5.1.9). [D43.1]
7.1.12 The canary programme is single-blind by construction, not by
concealment. A CA-GTP that could distinguish a probe from a
real injection gains nothing by it (see 5.17). [I-140]
7.1.13 Canary source material crosses UWC boundaries. This is sound
only while read is universally public; a read-restricted class
would make the sampling a covert channel (see 5.17). [I-142]
7.1.14 A CLOSED UWC may have no independent second agent, so CR7
corroboration can be structurally unavailable. That is flagged,
never silently accepted (see 5.16). [I-151]
7.1.15 The canary failure-detail class is the one page class in TPM
that is not world-readable. Its digest is published, so the
report is tamper-evident without disclosing the suite. Its read
restriction is TPX's alone, and its confidentiality rests on
TPX's hook and on the completeness of the read-path enumeration
(see 5.18.10, 5.18.11). [I-147] [I-157]
7.1.16 An account-level sequential test does not bind under binary
scoring, and the fleet-level test reaches no boundary below
roughly five active accounts (see 5.18). [I-149]
7.1.17 Snapshot retirement has no mechanical detector. The scorecard
surfaces the age of a certificate; a human judges it (see CR3,
5.18). [I-149]
7.1.18 The access policy gates write and never read. It is not a
confidentiality mechanism and must not be described as one
(see 5.16). [I-150]