Skip to content
Open
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
19 changes: 18 additions & 1 deletion docs/ci-verification.md
Original file line number Diff line number Diff line change
@@ -1,8 +1,25 @@
# CI verification

## Current selected-family replay: #6030
## Current selected-family replay: #7000 with #7001

The [selected record](../certificates/selected-result.json) binds conditional
κ = 0.00363005817 to the [application review](research/pr-intake-evidence/pr7000-application/PROOF-REVIEW.md),
[validation receipts](research/pr-intake-evidence/pr7000-application/VALIDATION.json),
and [explicit source closure](../certificates/frontier7000-source.json).
`make frontier7000-verify` follows that exact-source selected replay. The review
reuses eleven authenticated preceding executions and records seven fresh serial
application, scalar and complete native-catalogue executions. The earlier
monolithic timeout is retained separately and is not completed coverage.
Consult the receipts for exact commands, source heads and limits.

The five inherited framework assumptions remain. No complete uniform word,
literal multiplier execution, independent human proof or formal-kernel proof
is asserted. Routine research submissions do not need to rerun this
selected-family replay. The sections below preserve historical checkpoints.

## Historical selected-family replay: #6030

The [historical checkpoint](../certificates/checkpoints/frontier6030.json) binds conditional
κ = 0.003125804 to the [proof review](research/pr-intake-evidence/pr6030-audit/PROOF-REVIEW.md),
[validation receipts](research/pr-intake-evidence/pr6030-audit/VALIDATION.json),
and [explicit source closure](../certificates/frontier6030-source.json).
Expand Down
159 changes: 159 additions & 0 deletions research/fused-negative-half-tape/README.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,159 @@
# Fuse negation into the marked-tape negative half

A fixed reversible table computes and reverses the negative stochastic half
in at most **`4W+1` local transitions per direction** for every signed width
`W >= 2`. The bound is attained at `x=0` with either resident coin value.
This is a scoped written proof with exact finite corroboration. It reduces
the `6W+2` worst case reported by ToikY's #6809 under the same six-glyph tape,
parity, coin and head-return contract. It uses 35 coin-expanded rows per
direction. [Prior negative-half construction and transition bound][negative]

## Contract

Write `x=2q+p` with `q=floor(x/2)` and `p` in `{0,1}`. For a resident
read-only coin `u`, zxv's #6092 convention gives

```text
S(n,2,u) = floor(n/2) + [u < n mod 2]
S(-x,2,u) = -q-pu = ~q + (1-pu)
```

Here `~q=-q-1` is the signed two's-complement complement. The work head
starts at the clean parity marker in `P0 b0 ... b_(W-2) S(b_(W-1))`, with
the value in little-endian order. The forward call writes `S(-x,2,u)` in
the same layout and retains `Pp`. The inverse restores the original
`x` and `P0`. Both heads return. The work interval is exactly `W+1` cells
including parity, plus one stationary coin cell on a second tape. The
table uses the glyphs `0,1,P0,P1,S0,S1` and no additional working cells.

The directional entry and return states are `F0/FH` and `I0/IH`. Each
row reads one work glyph and the stationary coin, writes one work glyph
and moves the work head by at most one cell. The table supplies each row
explicitly, with `*` expanding to the two coin values. The coin is read
only. All calls stay inside the declared interval, so arbitrary exterior
cells retain their contents. [Literal table](TABLE.txt)

## Forward proof

Follow #6801's marked right shift. Scan to the marked sign, store its
original bit in control and write the complementary sign bit. On the
leftward pass each interior cell receives the complement of the carried
bit from its right. Its original bit becomes the new carry. Thus the
pass writes `~q` across all `W` value cells. Upon reaching the parity
marker, the carried original low bit `p` is written into `Pp`.
[Positive marked-shift construction][positive]

When `e=1-pu` is one, the table ripples an increment through the low
ones of `~q`. It visits the first zero or the marked sign, whichever
comes first, flips that bit and returns over the zeros it just wrote.
When `e=0`, the parity write also halts. The result is
`~q+e=-q-pu`. Both `~q` and `~q+e` fit in signed width `W`, since
`-2^(W-2) <= q <= 2^(W-2)-1`. The ripple agrees with signed addition,
including the all-ones word becoming zero.

## Inverse proof

On a canonical inverse entrance `(y,Pp,u)`, the table uses the same
`e=1-pu`. When enabled, it decrements `y` modulo `2^W` and returns over
the ones it just wrote. At parity it puts `p` into control, clears
the marker to `P0` and starts a rightward shift. In state `IUc`, the
incoming bit `c` is the original bit to write in the current cell. The
complement of the scanned ordinary bit becomes the next carry. At
the marked sign, the table accepts only `S(1-c)` and writes `S(c)`.
The leftward return only moves the head.

The sign rule requires that the two highest bits of the recovered
`q` agree. That is exactly the signed range condition
`-2^(W-2) <= q <= 2^(W-2)-1` for reconstructing `x=2q+p`.
For an integer `y` at the inverse boundary, the candidate is
`q=-(y-e)-1=-y-pu`. The only signed wrap in computing `y-e` occurs
at signed minimum with `e=1`. The wrapped word is signed maximum,
whose two highest bits differ. Its complement still has unequal
highest bits, so the guard rejects, as required by the integer
range test. Thus the inverse halts successfully exactly on the forward
image, where it restores `x=-2y+p(1-2u)` and `P0`. Malformed tapes and
rollback of rejected inputs are outside the interface.

## Local reversibility and exact time

For each target state, all incoming transitions in the table have the same
move. At fixed coin, the written glyph identifies the source row for that
target state. For example, `FD0` and `FD1` each receive a marked sign
from the scan or one of two distinct ordinary bits from the shift.
Each `IU` state receives `P0` from its unique entry branch at that
coin or one of two distinct ordinary bits while shifting. Ripple turns
and return loops have distinct written glyphs, and the halting branches
write distinct parity glyphs at fixed coin. These finite conditions make
the microstep injective. The target state recovers the predecessor head
position and the written glyph recovers the row. The verifier checks both
conditions on every row.

The shift and parity handling cost `2W+1` transitions per direction.
For `e=1`, let `k` be the index of the first zero of `~q`, capped at
`W-1`, using `W-1` if every bit is one. The carry phase adds exactly
`2(k+1)` transitions. The inverse decrement turns at the same cell
because the forward increment replaces exactly those low ones by zeros
and flips the turning bit. The all-ones case turns at the sign in both
directions. Therefore the exact invoice on a valid pair is

| Condition | Transitions per direction |
| --- | ---: |
| `e=0` | `2W+1` |
| `e=1` | `2W+2k+3 <= 4W+1` |

At `x=0` the complement is all ones and `e=1` for either coin, so the
table reaches `4W+1` in both directions. Relative to #6809's exact
worst case this saves `2W+1` transitions. The comparison is between
the two tables' worst cases. Optimality among all machines remains open.

## Reproduction

Run from the repository root with CPython 3.12. The verifier reads the
new table, prints a generated result and writes no files.

```sh
python3.12 -I -W error -B research/fused-negative-half-tape/verify.py
python3.12 -I -O -W error -B research/fused-negative-half-tape/verify.py
python3.12 -B scripts/research_package.py --package research/fused-negative-half-tape
python3.12 -B scripts/check_research_commit.py --package research/fused-negative-half-tape --check-source-history
```

The direct verifier reports `PASS_FUSED_NEGATIVE_HALF_TAPE`. It checks
4,088 exhaustive round trips at widths 2 through 10, 240 translated
boundary round trips through width 4,096, and the full inverse boundary
through width eight (1,016 images and 1,016 nonimages). It also checks five
rejecting controls for coin polarity, the inverse sign guard, an exterior
head movement and both local reversibility conditions. These runs execute
only the new table. The universal quantifier comes from the shift, ripple
and range arguments above.

The verifier caps the table input at 4,096 ASCII characters and the
simulator at `4W+1` transitions for widths through 4,096. The exhaustive
test domains and mutation set are fixed. Python allocates bounded tape
copies and table entries dynamically for verification. These data
structures describe the experiment and are outside the proved work-tape
budget.

## Open next questions

Bind one reached negative half to the marked layout. Start by pinning one
caller site and its signed width, then charge marker setup, resident coin
delivery, parity ownership and the return-state handshake. The supplied
markers, nonbinary glyphs and stationary coin are premises of this local
model. Physical encoding, repeated-call sequencing, reached precision,
native MOVE and the complete supplier bill remain open. This result
establishes a local transition saving and makes no κ claim.

## Credit

This contribution was produced by Nick Petalas. Credit ToikY's #6809 for
the negative marked-tape baseline and its open transition-constant question,
Nick Petalas's #6801 for the positive marked shift and local tape model, and
zxv's #6092 and #6066 for the signed arithmetic and restoration contract.
JoyboyBrian's #6095, AAAmirinu's #6244, Nick Petalas's #6026, Ali Romman's
#5599 and the retained stochastic and native lineage supply context. Exact
canonical provenance is recorded in [SOURCES.json](SOURCES.json).
Independent human and formal proof-kernel review are unclaimed. Apache-2.0.

[negative]: https://github.com/CrocSwap/integer-mult-bounds/pull/6809
[positive]: https://github.com/CrocSwap/integer-mult-bounds/pull/6801
36 changes: 36 additions & 0 deletions research/fused-negative-half-tape/SOURCES.json
Original file line number Diff line number Diff line change
@@ -0,0 +1,36 @@
{
"schema_version": 1,
"base_commit": "e54523517d0f03e58352e647c403a5af3e525a57",
"files": [
{
"path": "certificates/selected-result.json",
"usage": "provenance",
"git_blob": "75d77b328ea02ded535e7f23aee5b1e55d0bb1a8",
"sha256": "154c3d30455d778fd3343b6b4851f0fa4665f0f7625b0adb3675e2bafbbf6bf1"
},
{
"path": "docs/research/agent-task-catalog.json",
"usage": "provenance",
"git_blob": "3404fbdd5e7529a0674eef26114b5071e4586cec",
"sha256": "ead6cdc9d0af6d58e8a49744d8ec8c7c15e13d2ba93c276c434bcf9b51b93a81"
},
{
"path": "research/linear-negative-half-tape/PROOF.md",
"usage": "provenance",
"git_blob": "60ccb85519c5d9e9fc8f3cfa225eaeffb54361c2",
"sha256": "7e30db75ce6650d68c60bd753152502772c59c5f1755fb05852151c858b2a8d6"
},
{
"path": "research/linear-positive-half-tape/PROOF.md",
"usage": "provenance",
"git_blob": "f9d159fbb210ba89ac4f6541ac57ff658c82d5cc",
"sha256": "63eb48517f39cd0e66b23705939d74d2455a0c6b91f4da333ef768330b14e793"
},
{
"path": "research/reversible-signed-half-kernel/PROOF.md",
"usage": "provenance",
"git_blob": "93ebbe2a1eb08e6f2ee55d2b471b5377c154d4e4",
"sha256": "64c9b4b9f58639133f087d77f154b91026a18738bfb0830b93e96d5a30df9932"
}
]
}
41 changes: 41 additions & 0 deletions research/fused-negative-half-tape/TABLE.txt
Original file line number Diff line number Diff line change
@@ -0,0 +1,41 @@
# State read coin next write move
# * expands to one row for each of the two resident coin values.
F0 P0 * FS P0 1
FS 0 * FS 0 1
FS 1 * FS 1 1
FS S0 * FD0 S1 -1
FS S1 * FD1 S0 -1
FD0 0 * FD0 1 -1
FD0 1 * FD1 1 -1
FD1 0 * FD0 0 -1
FD1 1 * FD1 0 -1
FD0 P0 * FI P0 1
FD1 P0 0 FI P1 1
FD1 P0 1 FH P1 0
FI 0 * FR 1 -1
FI 1 * FI 0 1
FI S0 * FR S1 -1
FI S1 * FR S0 -1
FR 0 * FR 0 -1
FR P0 * FH P0 0
FR P1 0 FH P1 0

I0 P0 * ID P0 1
I0 P1 0 ID P1 1
I0 P1 1 IU1 P0 1
ID 0 * ID 1 1
ID 1 * IR 0 -1
ID S0 * IR S1 -1
ID S1 * IR S0 -1
IR 1 * IR 1 -1
IR P0 * IU0 P0 1
IR P1 0 IU1 P0 1
IU0 0 * IU1 0 1
IU0 1 * IU0 0 1
IU1 0 * IU1 1 1
IU1 1 * IU0 1 1
IU0 S1 * IB S0 -1
IU1 S0 * IB S1 -1
IB 0 * IB 0 -1
IB 1 * IB 1 -1
IB P0 * IH P0 0
84 changes: 84 additions & 0 deletions research/fused-negative-half-tape/TASK.json
Original file line number Diff line number Diff line change
@@ -0,0 +1,84 @@
{
"schema_version": 1,
"task_id": "joint-query-controller",
"priority_revision": 8,
"base_commit": "e54523517d0f03e58352e647c403a5af3e525a57",
"original_card": {
"id": "joint-query-controller",
"question": "Can a changed query demand or caller remove real work while satisfying native chronology and restoration?",
"starting_points": [
"docs/research/current-priorities.md",
"docs/research/research-family-status.md",
"certificates/selected-result.json",
"docs/research/pr-intake-evidence/pr5599-completion/PROOF-REVIEW.md"
],
"methods": [
"Construct and independently check one source-bound instance of the stated first experiment.",
"Seek an exact counterexample or target-specific complete-budget obstruction to that same proposed change."
],
"acceptance_gate": "A source-bound arbitrary-input identity and legal producer/consumer boundary with the complete paid and productive target invoice, including restoration, ownership, native alignment and extra charges. Price under the actual current model; historical ceiling.py assumptions must be established before use. A favorable optimistic relaxation keeps a route open but is not a completed supplier. A scoped local result remains useful; expansion requires the family-level budget gate.",
"first_step": {
"command": [
"python3",
"-B",
"scripts/research_selected_baseline.py"
],
"expected": "PASS_RESEARCH_BASELINE_BINDINGS_ONLY",
"inspect": "Read revision 8, the selected review and live exact-source handoffs. Then follow this card's bounded first_experiment.",
"scope": "Source-bound orientation only, not a supplier replay or theorem proof."
},
"first_experiment": "Choose an exact caller and a new demand/interface hypothesis. Read the scoped #5617/#5631/#5649/#5747 and #5720/#5745 objections before rebuilding an excluded schedule. Compare the complete matched source/consumer protocols including all retained helper outputs. For an active-mode injection, first check #5745: its canonical delivery claim repays the eight removed centre jobs. Audit that exact claim or change its channel/feedback hypothesis; do not repeat the same free-injection search. Stop if the full favorable budget fails or the required restored output is unobservable.",
"bottleneck_link": "Judge the actual current supplier contract and complete signed target/paid bill; frozen-profile ceilings require an applicability proof.",
"allowed_variations": [
"One exact source head, helper/cut region or quantified proposition and changed hypothesis; check live handoffs first. Different seeds are not different scopes."
],
"scope_change_triggers": [
"Changing source head, helper/cut domain, operation alphabet or caller boundary without recording the new scope.",
"Treating an archived proposal or local numerical pass as complete supplier admission."
],
"overlap_terms": [
"query demand",
"active mode",
"5720",
"5745",
"5747",
"caller"
],
"existing_checks": [
{
"scope": "PR #4166: author-checked installed producer/service/return/bridge connection; native-component comparison loses.",
"evidence": "research/qq16-caller-install/README.md",
"remaining": "Complete matched physical protocol bill, or changed source-compatible demand/circuit; no affordable supplier yet.",
"status": "author-checked",
"source_commit": "0b082d22a0e9a442d264ab0633e35ef26c790630",
"selected_record_sha256": "bb2eaa58647b353f88d0aec2eafb9d30df1be53589231b23e12ceef3e4bad129"
},
{
"scope": "Selected PR5599: complete helper-subspace construction and independent caller replay; five framework assumptions remain.",
"evidence": "docs/research/pr-intake-evidence/pr5599-completion/PROOF-REVIEW.md",
"remaining": "The next experiment must identify uncovered source-specific work. Successor archive intake does not transfer this admission.",
"status": "reviewed",
"source_commit": "09dd0965711ca7eeb169057a420abab1c9052137",
"selected_record_sha256": "2a0220e866417774606e0f991e41b9c1d26b1f43b97857238cf478dbc7d330b5"
}
]
},
"scope_status": "changed",
"actual_scope": "Lower the negative stochastic half's exact worst-case transition count on the supplied six-glyph marked tape.",
"scope_change_reason": "PR #6809 leaves the local transition constant open. This is one arithmetic component outside the full caller and query-scheduling gate.",
"bottleneck_link": "The fused shift removes an entire negation scan under the same local contract. Marker, coin, native and caller costs remain open, so the local saving does not establish a supplier improvement.",
"budget_minutes": 60,
"stop_rule": "Stop at the first faster reversible table with a written all-width proof and exact finite replay.",
"outcome": "partial",
"evidence": "README.md gives the exact 4W+1 worst case per direction. TABLE.txt and verify.py check the literal 70-row table, arithmetic endpoints, inverse image, time and failure controls.",
"author": "Nick Petalas",
"ai_assistance": "Independent human and formal proof-kernel review are unclaimed.",
"open_questions": [
{
"question": "Can one reached negative half use this marked-tape interface at a lower fully charged cost?",
"why_it_matters": "A local transition saving affects a supplier only after the actual producer and consumer can use it.",
"first_step": "Pin one reached site and budget marker setup, the signed width, resident coin delivery, parity ownership and the return-state handshake.",
"blocked_by": "A source-bound caller with a charged physical encoding and delivery contract."
}
]
}
34 changes: 34 additions & 0 deletions research/fused-negative-half-tape/VERIFY.json
Original file line number Diff line number Diff line change
@@ -0,0 +1,34 @@
{
"schema_version": 1,
"status": "ready",
"python": "3.12",
"requirements": [],
"inputs": [],
"commands": [
{
"argv": [
"{python}",
"-I",
"-W",
"error",
"-B",
"{package}/verify.py"
],
"timeout_seconds": 120,
"checks": "Literal table local reversibility, exact signed endpoints, head returns, transition count, inverse image and five failure controls."
},
{
"argv": [
"{python}",
"-I",
"-O",
"-W",
"error",
"-B",
"{package}/verify.py"
],
"timeout_seconds": 120,
"checks": "Literal table local reversibility, exact signed endpoints, head returns, transition count, inverse image and five failure controls."
}
]
}
Loading