Tighten README claims and distinguish current coverage from examples

Describe archive contents and origins directly, attribute physical retention, bring browsing before photograph details, and clarify current versus historical populations. Consolidate repeated limitations while preserving evidence scopes. Update the existing origin wording test and changelog.

Validation: 14 semantic consistency tests passed; current and pinned historical counts checked; README links resolve; dataset card configuration, generated workflow section, and loading code unchanged.

CHANGELOG.md

@@ -1,5 +1,16 @@
 # Changelog
 
+## October 24 README clarity
+
+- Describe recorded contents and collection origins directly; attribute physical
+  retention to the creator and preserve each relationship's evidence scope.
+- Bring collection selection and browsing ahead of photograph details, linking
+  implementation mechanics to the existing projection guide.
+- Distinguish the current 431-of-436 display-selection coverage from the retained
+  431-work selection table and the older 841-row loading example.
+- Consolidate repeated limits and remove the speculative annotation-study sentence.
+  Media, indexes, identities, selections, and historical receipts are unchanged.
+
 ## October 23 coordinated ingest
 
 - Add 51 unchanged MP4s and five confirmed work identities, with completeness unknown.

README.md

@@ -718,22 +718,19 @@ nine Pre-Standard dates remain unknown. Standardized acquisition began July 13,
 2025. The time-lapse indexes contain 2,400,224,173,549 indexed video bytes
 (approximately 2.40 TB). The real-time cohort adds 62,641,549,258 bytes.
 
-The recordings began as personal practice documentation and a durable record of
-manual work. The archive was initially organized as a dataset retrospectively;
-the current acquisition workflow is maintained prospectively. Recording was not
-initiated as a prospective scientific study. Each downstream study must state
-its own design and evidence requirements.
+The recordings began as personal documentation of drawing practice. Earlier
+recordings were organized into the dataset afterward; new recordings follow the
+documented acquisition workflow.
 
 Video is the primary published evidence for recorded process. The archive follows one creator working
 with physical drawing tools and materials on paper and illustration board.
 Metadata organizes the evidence and preserves unknowns. Here, **provenance**
 means documented origins and processing of source recordings, files, and
 derivatives, plus the evidence for links to physical works. It does not
-describe an ownership or custody history for the drawings. A populated
-`work_id` records creator-confirmed recording-to-work membership;
-physical-reference links retain separate evidence statuses.
+describe an ownership or custody history for the drawings.
 
-The archive documents recorded drawing processes and photographed drawing states. Their documented relationships support comparison while leaving intention, perceptual experience and unrecorded activity open to investigation.
+The archive preserves recorded drawing activity, photographs of physical drawing
+states, and evidence documenting their relationships.
 
 ## Choose a collection
 
@@ -768,81 +765,6 @@ The separate [2012–2016 archival holding](Earlier_Archive_2012_2016/README.md)
 has a public file manifest, but its media is held locally and excluded from these
 totals. Its historical scope label does not date every listed file.
 
-### Photographs and work images
-
-These three roles describe different selections and evidence scopes:
-
-| Role | Where to find it | Meaning |
-| --- | --- | --- |
-| Selected display photograph | [Drawing-photograph dataset](https://huggingface.co/datasets/maxwellinked/time-lapse-artifacts-drawing-photographs) and `recording_physical_references` | The front photograph currently selected for browsing a registered drawing. |
-| Selected high-resolution reference image | [`Work_Images`](Work_Images/README.md) | At most one active, authority-selected `high_res_primary` per work, with an exact `low_res_display` derivative. This is a selection role; suitability depends on documented properties and limitations. |
-| Historical photographic evidence | [`Physical_References`](Physical_References/Photographs/README.md) and its confirmation records | Preserved photographs and relationships interpreted at their documented revisions and evidence scopes. |
-
-**The first `Work_Images` ingest covers one work, `tlaw-794f3c7a1e5a05b6`.**
-Publishing that primary and its derivative does not automatically update the
-drawing-photograph dataset, recording projection, or separate GitHub browser.
-Their selections and publication revisions must be checked independently.
-`high_res_primary` does not supersede video for temporal claims; the historical
-`primary_inspection` purpose remains a separate label.
-
-Photograph identity and display selection answer separate questions: whether a
-particular photograph depicts a work, and why that photograph was chosen for
-browsing. An identity confirmation does not itself authorize a display choice.
-For a derivative, the evidence must also connect its exact display bytes to the
-confirmed photograph. Historical receipts retain their original claims.
-
-### Recording and photograph projection
-
-`recording_physical_references/archive` contains the same 1,384 indexed
-time-lapse recordings in Standard then Pre-Standard order. Its single `image`
-column preserves the selected display photograph for 1,205 Standard rows. The
-51 new Standard rows have registered work links but no selected photograph;
-they retain true null images. The 128 Pre-Standard rows also retain true null
-images because they have no registered work link. This view adds no recordings
-or work identities. Neither a photograph nor its selection establishes a
-recording's endpoint, timing, completeness, finish state, or frame alignment.
-
-The following counts describe the **prior photograph selection retained in the
-current projection**, using distinct Standard `work_id` values at revision
-`6fed0d073bdc307cdb81ea1dd8d4db6b1ecb04c7`:
-
-| Property | Work count | Scope |
-| --- | ---: | --- |
-| Embedded display photograph | 431 / 431 | One selected display per work; reused for its recording rows. |
-| Verified public source-photograph URL | 355 / 431 | Selected source hash and path match the pinned public source manifest. The other 76 works still have embedded displays. |
-| Documented selected-source resolution | 83 native; 272 limited; 76 unclassified in this projection | Source classifications, separate from display dimensions and suitability for a particular use. |
-
-Display JPEGs are at most 1024 pixels on the long edge and 512 KiB. Source and
-display hashes are distinct. Display dimensions do not establish original-source
-resolution, and an embedded display does not establish public source availability.
-These counts do not certify archive-wide hosted rendering or download success;
-report such checks separately with their revision, date, and tested population.
-
-The [scalar CSV](recording_physical_references.csv) contains the same recording
-rows and all 17 scalar fields, omitting only embedded images. Its explicit `\N`
-token means null, distinct from an empty work-ID string. Use `work_id` together
-with `source_photograph_sha256` and `display_image_sha256` to identify the exact
-selected photograph and derivative.
-
-The [projection manifest](viewer/recording_physical_references.manifest.json)
-binds the selection inputs, registry, and public exports. Its
-`status_at_generation` and the [validation receipt](viewer/recording_physical_references.validation.json)
-describe preparation-time validation. Repository publication is established by
-the containing immutable revision; hosted Viewer delivery requires separate checks.
-The manifest's `public_evidence_references.by_work_id` gives each current work
-separate routes for photograph identity, display selection, and source-to-display
-processing. Each entry binds the exact selected source and display hashes and
-locates the applicable public receipt, or states that the underlying evidence is
-private or not reproduced publicly. JSON pointers identify the exact object in
-the linked receipt; a match confirmation is not presented as a display-selection
-instruction. The retained selection map and processing-receipt names identify
-local build inputs, not publicly inspectable evidence. See the [rebuild guide](viewer/SELECTED_FRONT_PROJECTION.md)
-for input access boundaries, generation commands, and migration history.
-
-Public photographs retain the existing CC BY-NC 4.0 license and photographed-state
-limitations. Source photographs and confirmation evidence retain their documented
-storage and access boundaries.
-
 <!-- BEGIN generated workflow-subsets -->
 ### Select a Standard recording workflow
 
@@ -909,9 +831,12 @@ Install the tested loading version:
 python -m pip install 'datasets==4.8.5'
 ```
 
-This pinned revision contains all five native-Parquet configurations. The example
-selects `standard_individual_recordings` (841 rows); the other recording choices
-are `canonical`, `pre_standard`, and `standard_earlier`.
+**Reproducible example using an earlier snapshot:** 841 individual-recording
+rows. Current collection counts appear above.
+
+The pinned revision contains all five native-Parquet configurations. The example
+selects `standard_individual_recordings`; the other recording choices are
+`canonical`, `pre_standard`, and `standard_earlier`.
 `recording_physical_references` additionally contains embedded display photographs:
 
 ```python
@@ -940,6 +865,70 @@ For citation, retain this exact revision and selected IDs. The DOI snapshots
 identify earlier releases, not the 1,333-file revision used above; see
 [citation and documentation](#citation-and-documentation).
 
+## Photographs and work images
+
+These three roles describe different selections and evidence scopes:
+
+| Role | Where to find it | Meaning |
+| --- | --- | --- |
+| Selected display photograph | [Drawing-photograph dataset](https://huggingface.co/datasets/maxwellinked/time-lapse-artifacts-drawing-photographs) and `recording_physical_references` | The front photograph currently selected for browsing a registered drawing. |
+| Selected high-resolution reference image | [`Work_Images`](Work_Images/README.md) | At most one active, authority-selected `high_res_primary` per work, with an exact `low_res_display` derivative. This is a selection role; suitability depends on documented properties and limitations. |
+| Historical photographic evidence | [`Physical_References`](Physical_References/Photographs/README.md) and its confirmation records | Preserved photographs and relationships interpreted at their documented revisions and evidence scopes. |
+
+**The first `Work_Images` ingest covers one work, `tlaw-794f3c7a1e5a05b6`.**
+Publishing that primary and its derivative does not automatically update the
+drawing-photograph dataset, recording projection, or separate GitHub browser.
+Their selections and publication revisions must be checked independently.
+`high_res_primary` does not supersede video for temporal claims; the historical
+`primary_inspection` purpose remains a separate label.
+
+Photograph identity and display selection answer separate questions: whether a
+particular photograph depicts a work, and why that photograph was chosen for
+browsing. An identity confirmation does not itself authorize a display choice.
+For a derivative, the evidence must also connect its exact display bytes to the
+confirmed photograph. Historical receipts retain their original claims.
+
+### Recording and photograph projection
+
+`recording_physical_references/archive` contains the same 1,384 indexed
+time-lapse recordings in Standard then Pre-Standard order. Its single `image`
+column preserves the selected display photograph for 1,205 Standard rows. The
+51 new Standard rows have registered work links but no selected photograph;
+they retain true null images. The 128 Pre-Standard rows also retain true null
+images because they have no registered work link. This view adds no recordings
+or work identities.
+
+At this revision, selected display photographs cover **431 of the 436 active
+registered Standard works**; five works have no selected display photograph.
+
+The table below describes the **431-work selection retained from the earlier
+snapshot** at revision `6fed0d073bdc307cdb81ea1dd8d4db6b1ecb04c7`.
+Its denominators refer to that selection, not the current 436-work registry:
+
+| Property | Work count | Scope |
+| --- | ---: | --- |
+| Embedded display photograph | 431 / 431 | One selected display per work; reused for its recording rows. |
+| Verified public source-photograph URL | 355 / 431 | Selected source hash and path match the pinned public source manifest. The other 76 works still have embedded displays. |
+| Documented selected-source resolution | 83 native; 272 limited; 76 unclassified in this projection | Source classifications, separate from display dimensions and suitability for a particular use. |
+
+Display JPEGs are at most 1024 pixels on the long edge and 512 KiB. Source and
+display hashes are distinct. Display dimensions do not establish original-source
+resolution, and an embedded display does not establish public source availability.
+These counts do not certify archive-wide hosted rendering or download success;
+report such checks separately with their revision, date, and tested population.
+
+The [scalar CSV](recording_physical_references.csv) provides the same recording
+rows without embedded images. The [projection manifest](viewer/recording_physical_references.manifest.json)
+links the exact selected source and display hashes to separate evidence for
+photograph identity, display selection, and derivation, with private or
+unpublished evidence identified explicitly. The [projection guide](viewer/SELECTED_FRONT_PROJECTION.md)
+explains how to follow those references, interpret nulls and validation status,
+and rebuild the exports.
+
+Public photographs retain the existing CC BY-NC 4.0 license and photographed-state
+limitations. Source photographs and confirmation evidence retain their documented
+storage and access boundaries.
+
 ## Recording and drawing identities
 
 A **`record_id`** identifies one published video object. The time-lapse indexes
@@ -969,66 +958,40 @@ original recording survives or that the entire drawing process was recorded.
 
 ### Physical-work correspondence
 
-The archive documents reviewed relationships between recorded drawing activity
-and associated surviving physical drawings retained by the creator.
-
-For records with assigned `work_id` values, these relationships are organized
-around established work identities; multiple recording files may concern the
-same drawing. Pre-Standard record-to-physical links are documented separately
-below.
-
 Recording-to-work membership and physical-reference relationships are documented
-separately. The work-reference status projection distinguishes direct owner
-confirmation, proposed relationships derived from previously owner-confirmed
-identity links, assistant-reviewed visual correspondence, and unresolved cases.
-Derived and visually supported relationships are not presented as new owner
-confirmations.
-
-These relationships connect observations of a recorded physical process with
-evidence of the associated surviving material artifact. They may support
-historical, process-oriented, material, or other inquiries without prescribing a
-single research purpose.
-
-The [physical reference photographs layer](Physical_References/Photographs/README.md)
-documents an eligible subset by existing work ID against a pinned registry
-revision. Its manifest reports archived-source coverage and relationship-evidence limits.
-The current selected-front projections may also use receipt-bound display images
-from the separate photograph dataset; those display bytes retain their own hashes.
-The separate [work-image layer](Work_Images/README.md) records work-level still
-selection authority; its authority events do not change a physical-reference
-relationship status.
+separately. The [work-reference status table](Physical_References/work_reference_status.csv)
+distinguishes direct owner confirmation, proposed relationships derived from
+previously owner-confirmed identity links, assistant-reviewed visual
+correspondence, and unresolved cases. Derived and visually supported
+relationships are not new owner confirmations. Its [manifest](Physical_References/WORK_REFERENCE_STATUS_MANIFEST.json),
+[coverage summary](Physical_References/WORK_REFERENCE_STATUS_COVERAGE.json),
+[status schema](Physical_References/WORK_REFERENCE_STATUS_SCHEMA.md), and
+[contract](Physical_References/WORK_REFERENCE_STATUS_MANIFEST_CONTRACT.md)
+identify the applicable registry revision and evidence scope.
 
 A same-artifact relationship does not by itself establish the state shown at the
 end of an individual recording, complete process coverage, geometric
-registration, or verification of an inscription. Those are separate claims
-requiring separate evidence.
-
-A separate [Pre-Standard record-to-physical reference table](Physical_References/Pre_Standard_Record_References/record_physical_links.csv)
-records 128 creator-supplied, scope-limited links at its pinned source revision.
-It distinguishes photograph-to-reviewed-frame matches, direct in-hand
-front-to-frame checks, and record-level physical references without a
-confirmed frame locator. The table does not assign `work_id` values or
-establish whole-recording membership, final physical state, finish dates, or
-process completeness. Private photographs and in-hand observations are not
-publicly reproducible; see the [schema](Physical_References/Pre_Standard_Record_References/SCHEMA.md)
+registration, or verification of an inscription. The
+[photograph layer](Physical_References/Photographs/README.md) preserves its
+archived sources and relationship evidence; work-image selection authority does
+not change those relationship statuses.
+
+The separate [Pre-Standard reference table](Physical_References/Pre_Standard_Record_References/record_physical_links.csv)
+records 128 creator-supplied, scope-limited links at its pinned source revision:
+photograph-to-reviewed-frame matches, direct in-hand front-to-frame checks, and
+record-level physical references without a confirmed frame locator. These
+links do not establish final physical state, finish dates, or process
+completeness. Private photographs and in-hand observations are not publicly
+reproducible; see the [schema](Physical_References/Pre_Standard_Record_References/SCHEMA.md)
 and [evidence statement](Physical_References/Pre_Standard_Record_References/CONFIRMATION.md)
-for the supported claim and its limits.
-
-See the [work registry](Work_Identity/README.md), the generated
-[work-reference status table](Physical_References/work_reference_status.csv),
-its [release manifest](Physical_References/WORK_REFERENCE_STATUS_MANIFEST.json),
-the [coverage summary](Physical_References/WORK_REFERENCE_STATUS_COVERAGE.json),
-the [status schema](Physical_References/WORK_REFERENCE_STATUS_SCHEMA.md), and
-the [manifest contract](Physical_References/WORK_REFERENCE_STATUS_MANIFEST_CONTRACT.md).
-The status release reflects the work registry revision named in its manifest;
-its counts should be read against that revision.
+for each supported claim and its limits.
 
 ## Reading dates and recording starts
 
 Canonical filename times are documented **finish labels**. Under
-`standard_individual_recordings`, every member shares the complete drawing/group's
-finish label; it is not an individual recording's start or stop. Historical
-finish-label scope may be unspecified.
+`standard_individual_recordings`, recordings within a documented group share
+the finish label assigned to that drawing or group; it is not an individual
+recording's start or stop. Historical finish-label scope may be unspecified.
 
 - `recording_start_local` is a reviewed original device-filename timestamp,
   interpreted in the declared timezone. It is not an independently calibrated
@@ -1046,14 +1009,16 @@ from a shared finish label or playback duration.
 
 ## Evidence and limits
 
-The single-creator archive supports examination of this documented practice;
-generalization to other artists or settings requires additional evidence.
+Recording did not begin as a scientific study. The documentation describes
+collection and evidence; suitability for a particular analysis requires
+separate assessment.
+
+The archive documents one creator's practice. Population generalization and
+task-specific performance require evidence beyond the collection description.
 Acquisition and documentation vary across workflows, and workspace appearance
-may provide visual shortcuts for models. Downstream task performance and
-population, clinical, biometric, authorship, or leakage-safe benchmark claims
-require their own validation.
+may provide visual shortcuts for models.
 
-The current source release contains no frame-level bounding boxes, segmentation masks, pose, hand, pen-tip, stroke, aesthetic, or style labels. Small annotation-feasibility studies may be developed as derived work; any resulting labels remain separate from the source archive. The archive lacks universal per-record
+The current source release contains no frame-level bounding boxes, segmentation masks, pose, hand, pen-tip, stroke, aesthetic, or style labels. The archive lacks universal per-record
 capture calibration, original per-frame timestamps, camera calibration, and
 pixel-to-physical mapping. The nominal 0.2-second protocol setting is not a
 confirmed per-record measurement. Temporal or physical-motion claims must be
@@ -1094,11 +1059,11 @@ model release. Derived proxies should remove audio by default while preserving
 archival sources.
 
 The creator confirms ownership of the published videos and depicted
-creator-owned drawings. Each source time-lapse recording corresponds to a
-physical drawing retained by the creator. These physical works provide a
-material reference for provenance and metadata verification where applicable.
-Neither the ownership confirmation nor the physical-drawing linkage
-independently grants rights in incidental third-party content. The dataset is
+creator-owned drawings, and reports retaining the physical drawings represented
+in the time-lapse archive. Recording-to-work and physical-reference links are
+documented with their respective evidence scopes. Neither ownership confirmation
+nor physical-drawing linkage independently grants rights in incidental
+third-party content. The dataset is
 released under **CC BY-NC 4.0**, applying only to material for which the
 licensor holds the necessary rights. See the [license notice](LICENSE);
 commercial use outside its terms is not permitted.

tools/test_semantic_consistency.py

@@ -294,12 +294,15 @@ class SemanticConsistencyTests(unittest.TestCase):
         self.assertIn("- longitudinal-archive", readme)
         self.assertIn("- single-creator", readme)
         self.assertIn(
-            "The recordings began as personal practice documentation and a "
-            "durable record of manual work. The archive was initially organized "
-            "as a dataset retrospectively; the current acquisition workflow is "
-            "maintained prospectively. Recording was not initiated as a "
-            "prospective scientific study. Each downstream study must state "
-            "its own design and evidence requirements.",
+            "The recordings began as personal documentation of drawing practice. "
+            "Earlier recordings were organized into the dataset afterward; "
+            "new recordings follow the documented acquisition workflow.",
+            normalized,
+        )
+        self.assertIn("Recording did not begin as a scientific study.", normalized)
+        self.assertIn(
+            "The documentation describes collection and evidence; suitability "
+            "for a particular analysis requires separate assessment.",
             normalized,
         )