WEBVTT

1
00:00:00.000 --> 00:00:02.833
This overview explains how C21 compares

2
00:00:02.833 --> 00:00:10.467
seven methods under a common best-artifact contract
and C22 freezes a guarded source-packed path.

3
00:00:10.467 --> 00:00:18.033
A valid factorization is not enough when the task requires
the exhaustive-best artifact and complete charged work.

4
00:00:18.333 --> 00:00:23.667
Suppose two artifacts reconstruct the same target matrix exactly.

5
00:00:23.667 --> 00:00:28.500
Both satisfy a contract asking for any valid reconstruction.

6
00:00:28.500 --> 00:00:34.933
A stronger contract may require the unique winner under a frozen canonical ranking.

7
00:00:34.933 --> 00:00:39.167
For a simple illustration, rank by component count.

8
00:00:39.167 --> 00:00:45.667
A one-component artifact wins over a three-component artifact, even if both are exact.

9
00:00:45.667 --> 00:00:50.933
This toy field explains the distinction; the retained experiment uses its own complete

10
00:00:50.933 --> 00:00:56.433
frozen canonical ranking, which must be applied unchanged.

11
00:00:56.433 --> 00:01:02.233
Returning the larger artifact passes reconstruction while failing this stronger task.

12
00:01:02.233 --> 00:01:07.000
Correctness therefore includes more than a truth-vector match when the requested output

13
00:01:07.000 --> 00:01:10.200
specifies a particular best artifact.

14
00:01:10.200 --> 00:01:14.900
Compare methods under the same complete requirement.

15
00:01:14.933 --> 00:01:21.867
C21's seven-method table requires the exhaustive-best
GF(2) artifact under a common bounded task,

16
00:01:21.867 --> 00:01:25.200
rather than any valid factorization.

17
00:01:25.200 --> 00:01:32.033
Construction, proposal, completion, checks, cleanup
and wrappers belong to the charged request.

18
00:01:32.033 --> 00:01:37.167
Imagine two candidate artifacts that both reconstruct a truth vector.

19
00:01:37.167 --> 00:01:41.567
The frozen ranking selects the smaller canonical score.

20
00:01:41.567 --> 00:01:48.600
Returning the other valid artifact satisfies
reconstruction but misses the best-artifact contract.

21
00:01:48.600 --> 00:01:54.833
In C21, source-packed construction is also distinct from proposal success:

22
00:01:54.833 --> 00:02:03.200
the representation path can run even when its cut proposal
abstains and exact screened completion chooses the answer.

23
00:02:03.200 --> 00:02:08.400
The retained narrow source-packed
advantage leaves little routing headroom.

24
00:02:08.400 --> 00:02:13.833
C22 freezes an opt-in source-packed
screened path with exhaustive restoration,

25
00:02:13.833 --> 00:02:17.767
refusal fallback and shadow identity checks.

26
00:02:17.767 --> 00:02:23.967
That is implementation readiness, not fresh
timing evidence or production promotion.

27
00:02:24.267 --> 00:02:30.600
A source-packed representation can be constructed even when a later cut proposal abstains.

28
00:02:30.600 --> 00:02:34.433
Construction and proposal success are different stages.

29
00:02:34.433 --> 00:02:40.567
A refusal to propose does not imply that the representation was never built.

30
00:02:40.567 --> 00:02:45.700
The route can continue with exact screened completion, choose the required artifact,

31
00:02:45.700 --> 00:02:48.400
verify it, and deliver the answer.

32
00:02:48.400 --> 00:02:53.867
All these stages, including the unsuccessful proposal work, belong to the charged request

33
00:02:53.867 --> 00:02:56.900
when the caller pays for them.

34
00:02:56.900 --> 00:03:01.700
An opt-in shadow path can record candidate behavior while preserving the control answer

35
00:03:01.700 --> 00:03:04.600
and checking restoration and identity.

36
00:03:04.600 --> 00:03:08.533
That establishes operational checks within its scope.

37
00:03:08.533 --> 00:03:11.933
It is not a substitute for fresh timing evidence or permission to

38
00:03:11.933 --> 00:03:15.319
change a production default.

39
00:03:15.319 --> 00:03:15.333
change a production default.

40
00:03:15.333 --> 00:03:20.000
Separate construction, proposal, completion, and checking costs.

41
00:03:20.000 --> 00:03:26.526
Implementation readiness does not replace
fresh timing evidence or production promotion.