codawork update

Author: PeterHiggins19

drafts/codawork-2026/ADVERSARIAL_PANEL.md

@@ -47,7 +47,7 @@
 
 ---
 
-### DA-06 — "Why should we trust your 17-error catalogue? Where is the validation?"
+### DA-06 — "Why should we trust your 25-error catalogue? Where is the validation?"
 
 **Why they ask:** CoDa researchers value published literature and peer review.
 
@@ -76,7 +76,7 @@
 ### Q1 — "Isn't HUF just MEWMA with different branding?"
 **A:** "MEWMA stores energy. HUF-GOV is stateless. They are complementary, not redundant."
 
-### Q2 — "Why dual metrics? Isn't that mathematically inconsistent?"
+### Q2 — "Why three diagnostics? Isn't that mathematically inconsistent?"
 **A:** "It's diagnostic, not fusion. When TV and Aitchison disagree, that's information."
 
 ### Q3 — "Your instrument breaks at zero. Why should we trust it?"
@@ -97,9 +97,9 @@
 
 | Question | Answer |
 |----------|--------|
-| Why dual metrics? | Diagnostic disagreement signal. |
+| Why three diagnostics? | Diagnostic disagreement signal. |
 | Why open loop? | Prevents stored energy contamination. |
-| Why 17 errors? | Because the union inherits all errors from both sides plus new ones. |
+| Why 25 errors? | Because the union inherits all errors from both sides plus new ones. |
 | Why Ramsar? | Real ecological compositions; physical-world validation. |
 | Why not signed or complex CoDa now? | CONJ-001: too early; union must validate first. |
 | Why should CoDa care? | Because this is the first monitoring-oriented calibration study. |

drafts/codawork-2026/CODAWORK_PREP_GUIDE.md

@@ -61,7 +61,7 @@ d_A(x,y) = √(Σᵢ<ⱼ (ln(xᵢ/xⱼ) - ln(yᵢ/yⱼ))²) · (1/D)
 
 Or equivalently: the Euclidean distance between the CLR-transformed vectors.
 
-**Why it matters for you:** This is the "right" distance on the simplex according to CoDa theory. The conference demonstrator now computes both TV distance and Aitchison distance side-by-side as a dual-metric protocol. If someone asks "have you computed Aitchison distance?", the answer is: "Yes — we now run both TV and Aitchison on every observation. Agreement is treated as robustness; disagreement is treated as diagnostic information (dominant vs trace carrier movement). The open question is which metric should anchor the monitoring protocol."
+**Why it matters for you:** This is the "right" distance on the simplex according to CoDa theory. The conference demonstrator now computes both TV distance and Aitchison distance side-by-side as a three-diagnostic protocol. If someone asks "have you computed Aitchison distance?", the answer is: "Yes — we now run both TV and Aitchison on every observation. Agreement is treated as robustness; disagreement is treated as diagnostic information (dominant vs trace carrier movement). The open question is which metric should anchor the monitoring protocol."
 
 ### 4. The Zero Problem
 
@@ -111,7 +111,7 @@ The CoDa equivalent of PCA biplots. Based on CLR-transformed data. Shows the rel
    → The conference demonstrator now computes both. TV distance is retained for monitoring interpretability; Aitchison distance is the simplex-native geometric metric. Agreement between them is treated as robustness; disagreement is treated as diagnostic (dominant vs trace carrier movement). The open question is which should anchor the monitoring protocol — that's exactly why I'm here.
 
 2. **"How do you handle zeros in your compositions?"**
-   → Event-first protocol: a carrier reaching zero is treated first as a domain event (flag it, record the TV velocity spike), then CoDa zero-handling is applied only if further geometric analysis is required. Greenacre's chiPower is being explored as a post-event correction layer that preserves subcompositional coherence. Open to CoDa community guidance on best practice. This is documented as E-03/E-17 in the 17-error calibration catalogue.
+   → Event-first protocol: a carrier reaching zero is treated first as a domain event (flag it, record the TV velocity spike), then CoDa zero-handling is applied only if further geometric analysis is required. Greenacre's chiPower is being explored as a post-event correction layer that preserves subcompositional coherence. Open to CoDa community guidance on best practice. This is documented as E-03/E-17 in the 25-error calibration catalogue.
 
 3. **"Have you considered log-ratio transformations?"**
    → Yes. CLR and ILR are now implemented in the conference demonstrator alongside the original TV-based metrics. The conference posture is that Aitchison distance and log-ratio views are included for exploratory calibration — the open question is not whether they exist in the toolchain, but how they should anchor the monitoring protocol, what the correct null model is, and how zeros and subcompositions should be governed.

drafts/codawork-2026/COIMBRA_TRIM_LIST.md

@@ -19,7 +19,7 @@ Walk in narrow. Speak plainly. Show the case. Invite correction. Let CoDa do its
 | Item | Why |
 |------|-----|
 | **CoDaWork Packet v3** | Front-door document. Claim, methods challenge, EMBER case, metric correction appendix. Cleanest first-contact artifact. |
-| **Abstract v2** | 208 words. Union thesis, dual-metric protocol, 17-error calibration framing. Matches the packet. |
+| **Abstract v3 | 197 words. Three diagnostics, CR results, Backblaze cross-domain, 25-error calibration. Matches all conference artifacts. |
 | **One clean EMBER figure** | Stable total, shifting mix. Not the full analyzer — just the simplest defensible picture. |
 | **Spoken posture** | "Here is a compositional monitoring claim. Here is one public case. Here is how to break it." |
 

drafts/codawork-2026/ROOM_CONTROL.md

@@ -32,7 +32,7 @@ Engineers test you on architecture, failure modes, stored energy, signal flow, w
 Say: "HUF-GOV is a phase discriminator — stateless, open loop. HUF-CLS is the PLL — closed loop, stored energy. Only GOV belongs in scientific monitoring." This resolves 80% of engineering objections before they form.
 
 **Move E2 — Show the error catalogue**
-Engineers love failure modes. Say: "We mapped 17 error sources. Each has a detection test and a governance action."
+Engineers love failure modes. Say: "We mapped 25 error sources. Each has a detection test and a governance action."
 
 **Move E3 — Use their metaphors**
 Terms to use: "stored energy", "loop stability", "aliasing", "dimensionality change as a topology shift".
@@ -74,7 +74,7 @@ After a strong point, pause. Let them fill the silence with agreement.
 
 ## Three Hardest Scenarios
 
-**Scenario A — A mathematician challenges the dual metric**
+**Scenario A — A mathematician challenges the three-diagnostic protocol**
 You say: "Disagreement is a diagnostic signal. We don't fuse the metrics — we compare them."
 
 **Scenario B — An engineer insists HUF is a PLL**

drafts/codawork-2026/THE_LINEAGE.md

@@ -132,7 +132,7 @@ This is not an alliance between a framework and a theory. This is an alliance be
 
 That's what I'm here for. Not to present a framework. Not to defend an instrument. To put a screwdriver on the table next to a math book and see what we can calibrate together.
 
-*Mathematically: the HUF-CoDa union is a calibration study. CoDa provides the Aitchison geometry, the log-ratio transforms, the statistical optimization. HUF provides the monitoring doctrine, the governance architecture, the open-loop discipline, the audit trail. The union produces capabilities neither can build alone: a statistically rigorous, governancially sound, dimensionality-aware, dual-metric monitoring instrument for compositions in motion. The 17-error catalogue maps the failure modes. The calibration study tests the joint. The alliance reduces the mutual ignorance. The simplex is the same everywhere.*
+*Mathematically: the HUF-CoDa union is a calibration study. CoDa provides the Aitchison geometry, the log-ratio transforms, the statistical optimization. HUF provides the monitoring doctrine, the governance architecture, the open-loop discipline, the audit trail. The union produces capabilities neither can build alone: a statistically rigorous, governancially sound, dimensionality-aware, three-diagnostic monitoring instrument for compositions in motion. The 25-error catalogue maps the failure modes. The calibration study tests the joint. The alliance reduces the mutual ignorance. The simplex is the same everywhere.*
 
 ---
 

drafts/codawork-2026/WHAT_HUF_IS.md

@@ -160,7 +160,7 @@ Forty years of active filter crossover design. Six months of formalization. The
 
 The mathematics already exists. Aitchison formalized the simplex geometry in 1982. Egozcue and Pawlowsky-Glahn built the Hilbert space structure. The MEWMA-CoDa researchers built statistically optimal control charts. Filzmoser built robust outlier detection.
 
-The monitoring doctrine already exists. HUF's open-loop discipline, governance reference management, kill test, coherence chain, stored energy audit, and 17-error calibration study.
+The monitoring doctrine already exists. HUF's open-loop discipline, governance reference management, kill test, coherence chain, stored energy audit, and 25-error calibration study.
 
 What doesn't exist yet is the union. The instrument that combines CoDa's mathematical rigour with HUF's monitoring architecture. The instrument that is statistically optimal AND governancially sound. The instrument that watches compositions move and tells you when to care.