# Research Site brief and next-window prompt

Prepared September 29, 2026. Research checkpoint: **C1634 complete; C1635 not begun.**

## Purpose

Build a research Site within ChatGPT, complementary to Dean's main **490d.com**. Its purpose is to make the accumulated research navigable and help explain **how the chronological families fit together**. The guiding Strategy seeks a small, state-aware set of source inputs, operations and relations that reconstructs whole families. The Site should support that investigation while keeping the original source documents, dated research and current synthesis distinguishable.

Dean plans to provide the earlier archive and this companion archive together. Preserve them as separate, immutable source packages, with a combined searchable catalog. Use the supplied archive inventory for actual file counts: labels such as **C1431** and **C1634** are research-action identifiers, not file counts.

## Read first

1. This companion archive's `00_START_HERE.md` instructions and inventory, followed by the earlier archive's handoff.
2. `Source_Updates/Research_Strategy_v0_2.md`, read in full. Its dated immediate next-step section is historical; later completed research and author directions determine the current continuation.
3. The current `490d_Chronological_Families_Explanation_C1431.md` and `490d_Chronological_Families_Explanation_C1431_Print.pdf`. **The stable C1431 filename now contains the synthesis extended through C1634.** This is the default reading view. Earlier copies remain dated evidence.
4. `Strategy_Review_After_C1634.md` and the C1585–C1634 continuation for the present conclusions and unresolved question.
5. The relevant source/proof indexes, technical companion, reconstruction ledger, model files and frozen source snapshots when following a particular claim.

Where repeated filenames occur, use the companion inventory's explicit role and version information. Do not choose a file merely by its basename or by filesystem modification time. Preserve checksummed evidence copies; use separate portable reading copies for corrected local links or display formatting.

## Suggested first Site release

| View | Research purpose |
|---|---|
| Start / current explanation | Present the connected account, latest checkpoint, family map and printable PDF. |
| Chronological families | Navigate Regular, Cumulative, Rounded, inverse, calendar-Key, Covenant, counted/NT and other documented families; show their demonstrated connections. |
| Source library | Search and filter documents by source number, title, edition, manuscript tradition, mode and research range. |
| Claims and reconstructions | Show a claim's named source rows, active state, allowed operations, anchors, calculation, dependencies and evidence. |
| Research history | Browse dated continuations, questions, results, reassessments and verified completion checkpoints. |
| Open questions | Keep a short, current notebook of the next bounded questions and what would discriminate between proposed explanations. |

Begin with a working reading-and-search experience using the supplied corpus. Add interactive comparison views when their inputs and operations are already defined. For example, the three inverse spines can be compared side by side while preserving manuscript, mode, anchor stage, partition and endpoint role. A shared date can have several source roles; display those roles separately.

A claim record should carry a stable ID, short statement, source file and edition/hash, section or table locator, research action, selected state, input rows, operation and domain, exact output, dependencies, and claim status. Distinguish a supplied textual value, selected reconstruction, adopted anchor, generated coordinate, mathematical consequence and interpretation. Record failures and limited domains alongside successes. Several dependent interval displays can belong to one evidence family.

Cross-links should connect the current explanation to its proofs and original rows. Public 490d.com links can accompany local source snapshots; a changing public URL should not erase the exact edition used. Site-specific navigation, labels and derived indexes belong in their own layer, leaving frozen research records intact.

## Current source and scope decisions

- The latest, most thorough **File52c** supplied in this conversation controls this research phase. Retain its exact edition and hash. The older similarly named File52c in the initial source collection is historical.
- Retain the supplied **LXX reversed spines and accepted SP solution** study as the basis of the comparative extension. MT is primary, LXX supplies the selected secondary comparison, and the accepted SP Fall inputs remain explicit.
- Keep native LXX **Cainan ON / full 430** distinct from deliberately leveled comparisons. The selected LXX Lamech pair is **182/753**; **777** remains an appendix comparison and **188** is nonoperative. Do not restore older selections from the Strategy's historical discussion.
- Regular birth intervals and Cumulative lifespan sums are different measurements. Cumulative Noah **950** and Shem **600** must retain their roles; the Regular Flood uses Noah **600**. Gear exploration remains confined to the admitted Regular Noah/Shem/Flood domain.
- Decimal reversal acts **once on the declared original duration segments**, preserving trailing-zero placeholders as defined in File52c. Do not reverse BC date labels or assume reversal is additive.
- **Deeper mirror analysis and the inverse of the inverse remain deferred.** The specifically authorized 2700-year Rounded translation, including restoration **536 BC → Rounded AD 2166**, is part of the current result.
- Civil elapsed spans crossing the epoch use **BC + AD − 1**. File52c's Rounded coordinates use **q(BC B)=1−B**, **q(Rounded AD A)=A−1**, hence cross-side width **B+A−2**. Keep the convention attached to each calculation.
- The Strategy's providential interpretation remains its stated interpretive premise. Arithmetic compatibility, historical transmission claims and theological interpretation retain their distinct claim levels.

## What the latest work contributes

The current synthesis connects whole named paths instead of accumulating isolated attractive intervals. Located source-row changes explain the MT–LXX Cumulative Flood difference of **910**; the accepted SP lower leg and its one-pass reversal explain the MT–SP inverse difference of **2100**. Three **4900-year** placements connect by complete-interval translations of **910** and **600**.

The selected **12600** path has ordered components **5600+2100+4900**. Its established Covenant refinement retains the **11270** radius and the **7000** Flood interval. The authorized 2700-year Rounded translation preserves those components through **Rounded AD 2166**. The SP inverse Flood at **11206 BC** supplies **9800** to Conquest **1406 BC** and **11270** to civil **AD 65**. Each display should link to its source conditions and exact derivation.

The strongest next research question is source generation: can an already-fitted row relation be reduced further while retaining the named row positions and the Regular/Cumulative distinction? A new test must state allowed inputs and an output not used to choose the rule. If no discriminating test is available, improving the connected explanation remains useful work. Broad endpoint searches are not the next priority.

## Paste into the new window

> I am supplying my earlier 490d research archive and a companion archive of subsequent work, plus the original source files. Build a research Site within ChatGPT, complementary to my main 490d.com, using this material as the research basis.
>
> First read the companion 00_START_HERE.md instructions, archive inventory, this Site brief, the full Research Strategy, and the current Chronological Families Explanation. Preserve the separate source archives and create a combined catalog with source, edition, hash, research range and current/historical roles. Check the supplied inventory for actual file counts; C-series labels identify research actions.
>
> The current completed checkpoint is C1634; C1635 has not begun. The main explanation retains C1431 in its filename but has been extended through C1634. Use its latest copy as the default reader, with links to the technical evidence. The interrupted C1132–C1431 cycle was completed and independently recovered; use its verified recovered final package, not the stale working folder ending at C1291.
>
> The larger goal is to explain how the chronological families fit together. Build a clear, useful reading and research interface: current synthesis and PDF, family navigation, searchable sources, claims linked to exact source rows and proofs, research history, and a short open-questions notebook. Preserve named manuscript/mode states, operation domains, anchors, negative results and dependencies. Keep Site presentation edits separate from frozen evidence and canonical source documents.
>
> Use the latest File52c and the supplied accepted LXX/SP study. Deeper mirror analysis and inverse of the inverse remain deferred; the authorized 2700-year shift to Rounded AD 2166 stays included. Keep civil and Rounded epoch conventions distinct. Follow the current Sites workflow when implementing the Site, and preserve all supplied material and references needed to reproduce the research.