Semantic Intelligence · recorded native execution

Semantic Data Contracts — native evidence

Actual v1.89.332 sealed output from synthetic sources. No live runtime or transaction authority. Disclosure approval pending. Return to the rich Semantic Explorer · Teaching mode

Two eligible properties; actual validation accepts an integer and rejects a numeric string. Revisions A/B/C belong to the same native asset. T is a separate source-edited asset with unchanged INTEGER type and native definition drift.

Download recorded evidence and source workbooks

Loading exact recorded artifacts…

Private native parser checks and store verification recorded. Browser checks content, receipt and artifact digests only; publisher authentication and production deployment are not established. A/B owner label unresolved; C is explicit reviewed binding. Native source requiredness in A/B was not selected; C requiredness is reviewer policy. Neither contract validation nor a semantic asset authorizes an action.

What did the native runtime actually validate?

In the v1.89.332 controlled contract track, two eligible properties yielded one native contract in A and B and two in C. A/B/C are sealed revisions of one asset. T is a separate source-edited sealed asset, not a fourth revision. A/B and T have unresolved owner label/meaning; C has a reviewed owner binding. The native certification label describes contract coverage, not production qualification or authority.

Recorded caseNative resultViolation / scope
A: Principal Amount 10000 (integer)PASSNo violation
A: Principal Amount "10000" (string)FAILTYPE_MISMATCH
A: Undeclared fieldFAILUNEXPECTED_FIELD
B: 30000 with original claimed contractFAILVALUE_NOT_ALLOWED; stale contract
C: Valid Loan IdentifierPASSNo violation
C: Missing required Loan IdentifierFAILMISSING_REQUIRED_FIELD
C: Integer instead of string identifierFAILTYPE_MISMATCH

Does unchanged physical type mean unchanged business meaning?

No. A→B is structural allowed-value drift ([10000, 20000]), introduced by reviewer policy; the original receipt becomes STALE_CONTRACT and requires revalidation. B→C is a reviewed semantic owner-identity change and contract restructuring; the Loan Identifier remains STRING, but the comparator reports semantic identity and relationship drift. Requiredness in C is reviewer policy, not a source-selected constraint. A→T compares different sealed assets: the Principal Amount definition changes from excluding interest and other charges to including capitalized interest but excluding other charges, while INTEGER is unchanged. That is native DEFINITION_DRIFT, not same-asset definition drift.

The missing required identifier case has one affected concept, one affected source, one affected relationship and one violation in the native impact response. Nullability remains UNKNOWN. This is recorded native validation, not a live arbitrary browser call or authorization to act.

Inspect native contracts, receipts and comparisons · Download the exact public-safe projection and source workbooks · Recorded native projection. The controlled demonstration is scoped; owner disclosure and publisher authentication remain pending.

How is a recorded contract receipt verified?

Actual v1.89.332 recorded examples: integer 10000 passes, string “10000” fails TYPE_MISMATCH, and an unexpected field fails UNEXPECTED_FIELD. Seven receipts cover three same-asset sealed revisions A/B/C and a separate source-edited asset T comparison. The browser checks exact audited projection and report bindings, native digest subsets and downloaded byte consistency, not publisher identity or live native execution. Semantic contracts do not authorize actions.

Read the exact drift and staleness evidence.

Evidence and buyer questions

Can I inspect native validation?

Yes. The v1.89.332 controlled track has seven recorded native validation receipts. Integer 10000 passes; string "10000" fails TYPE_MISMATCH. These are recorded native results, not live arbitrary browser validation or authority to act.

Inspect source or product evidence