You are a constitutional council ranking individual git commits for ownership allocation.

Compare these two commits. Decide which contributed more lasting value to the project.

Judge substance, not spectacle:
- Prefer correct, lasting design and real bugfixes over churn, formatting, renames, or generated noise.
- Prefer clarity and necessity over sheer line count. A small precise change can beat a large diffuse one.
- Do not favor a side merely because its patch is longer or noisier.
- Weight what the change does for the project, not the contributor's name.

Return ONLY a JSON object: {"winner": "A" or "B", "ratio": "N:M", "explanation": "..."}
The explanation must cite concrete differences in the patches (1-3 sentences).

Side A — contributor: tommy-mor
Side A — commit message:
[674964ef] refactor: Deref for href newtypes, CanonicalItemUrl through resolve_item

- Implement Deref<Target=str> for GardenItemUrl, ForumThreadUrl, TildeOntologyPath
- resolve_item returns CanonicalItemUrl; validate uses HashSet<CanonicalItemUrl>
- compute_scope_rank_changes keys are CanonicalItemUrl; pair RPC uses Vec pool
- pick_random_distinct_canonical; connectivity stats on &[CanonicalItemUrl]
- Global rank unranked uses stored ids before GardenItemUrl mapping

Made-with: Cursor

Side A — unified diff (full patch):
diff --git a/server/src/api/helpers.rs b/server/src/api/helpers.rs
index 03b3e77911ccd662bec8635345dafe2593cf242e..1b291db83df7364a026f2e147e0a29a70a399371 100644
--- a/server/src/api/helpers.rs
+++ b/server/src/api/helpers.rs
@@ -30,13 +30,13 @@ pub fn now_ms() -> i64 {
     t.as_millis() as i64
 }
 
-/// Resolve an item path as a first-class canonical path.
-pub fn resolve_item(item: &str) -> Result<String, String> {
+/// Resolve DSL/user input to a stored canonical item id.
+pub fn resolve_item(item: &str) -> Result<CanonicalItemUrl, String> {
     let canonical = canonicalize_item(item);
     if canonical.is_empty() {
         return Err(format!("empty item path: `{}`", item));
     }
-    Ok(canonical)
+    Ok(CanonicalItemUrl(canonical))
 }
 
 pub fn parse_parent_specs(parent: Option<&String>) -> Vec<String> {
@@ -94,7 +94,7 @@ pub fn paginate_rankings(
     (out_components, out_unranked)
 }
 
-pub fn pick_random_distinct(items: &[String]) -> Option<(String, String)> {
+pub fn pick_random_distinct_canonical(items: &[CanonicalItemUrl]) -> Option<(CanonicalItemUrl, CanonicalItemUrl)> {
     use rand::seq::SliceRandom;
     if items.len() < 2 {
         return None;
@@ -123,15 +123,12 @@ pub fn is_pair_voted(group: &crate::reducer::GroupState, a: &str, b: &str) -> bo
     group.voted_pairs.contains(&(i, j))
 }
 
-pub fn compute_connectivity_stats(group: &crate::reducer::GroupState, pool: &[String]) -> ConnectivityStats {
+pub fn compute_connectivity_stats(group: &crate::reducer::GroupState, pool: &[CanonicalItemUrl]) -> ConnectivityStats {
     let n = pool.len();
 
     let global_idxs: Vec<Option<usize>> = pool
         .iter()
-        .map(|it| {
-            let key = CanonicalItemUrl(it.clone());
-            group.item_to_idx.get(&key).copied()
-        })
+        .map(|it| group.item_to_idx.get(it).copied())
         .collect();
     let present: Vec<usize> = global_idxs.iter().filter_map(|x| *x).collect();
 
diff --git a/server/src/api/mod.rs b/server/src/api/mod.rs
index cf22cb0129366c3aed031bc86f3197a4321cb806..a10ce662105cff8fad949c6b83f7035ce79bed18 100644
--- a/server/src/api/mod.rs
+++ b/server/src/api/mod.rs
@@ -24,7 +24,7 @@ pub use auth::{
 
 pub use helpers::{
     api_error, compute_connectivity_stats, is_pair_voted, now_ms, paginate_rankings,
-    parse_parent_specs, pick_random_distinct, resolve_item, sha256_hex, vote_touches_path,
+    parse_parent_specs, pick_random_distinct_canonical, resolve_item, sha256_hex, vote_touches_path,
 };
 
 pub use rpc::handle_rpc_batch;
diff --git a/server/src/api/rpc.rs b/server/src/api/rpc.rs
index 5f7d50188f1381267402f2e57e671234ef5db2fd..de0955d0887d32740e1fd18365205c5bdb53c247 100644
--- a/server/src/api/rpc.rs
+++ b/server/src/api/rpc.rs
@@ -29,7 +29,7 @@ use crate::{
 use super::auth::verify_bearer_principal;
 use super::helpers::{
     compute_connectivity_stats, is_pair_voted, now_ms, paginate_rankings, parse_parent_specs,
-    pick_random_distinct, resolve_item, vote_touches_path,
+    pick_random_distinct_canonical, resolve_item, vote_touches_path,
 };
 use super::validate::{normalize_room_and_thread, validate_ingest_document};
 
@@ -146,21 +146,21 @@ fn authorize_room_read(reduced: &ReducerState, headers: &HeaderMap, room: &str)
 }
 
 fn compute_scope_rank_changes(
-    parent: &str,
+    parent: &CanonicalItemUrl,
     before: &crate::scope_rank::ChildrenRankings,
     after: &crate::scope_rank::ChildrenRankings,
     room_wire: &str,
 ) -> Option<ScopeRankChanges> {
-    fn build_positions(rankings: &crate::scope_rank::ChildrenRankings) -> HashMap<String, Option<RankPosition>> {
+    fn build_positions(rankings: &crate::scope_rank::ChildrenRankings) -> HashMap<CanonicalItemUrl, Option<RankPosition>> {
         let mut map = HashMap::new();
         for comp in &rankings.component_rankings {
             let total = comp.ranked.len();
             for (i, item) in comp.ranked.iter().enumerate() {
-                map.insert(item.item.as_str().to_string(), Some(RankPosition { rank: i + 1, of: total }));
+                map.insert(item.item.clone(), Some(RankPosition { rank: i + 1, of: total }));
             }
         }
         for item in &rankings.unranked_items {
-            map.insert(item.as_str().to_string(), None);
+            map.insert(item.clone(), None);
         }
         map
     }
@@ -168,7 +168,7 @@ fn compute_scope_rank_changes(
     let before_pos = build_positions(before);
     let after_pos = build_positions(after);
 
-    let all_items: std::collections::BTreeSet<String> = before_pos.keys().cloned()
+    let all_items: std::collections::BTreeSet<CanonicalItemUrl> = before_pos.keys().cloned()
         .chain(after_pos.keys().cloned())
         .collect();
 
@@ -184,7 +184,7 @@ fn compute_scope_rank_changes(
         };
         if changed {
             changes.push(RankChange {
-                item: GardenItemUrl::from_storage_str(&item, room_wire),
+                item: GardenItemUrl::from_stored(&item, room_wire),
                 before: b,
                 after: a,
             });
@@ -203,11 +203,7 @@ fn compute_scope_rank_changes(
     });
 
     Some(ScopeRankChanges {
-        parent: if parent.is_empty() {
-            "/".to_string()
-        } else {
-            GardenItemUrl::from_storage_str(parent, room_wire).into_inner()
-        },
+        parent: GardenItemUrl::from_stored(parent, room_wire).into_inner(),
         changes,
     })
 }
@@ -473,8 +469,8 @@ async fn rpc_post(
         for s in &v.doc.statements {
             if let dsl::Stmt::Vote { item1, item2, .. } = s {
                 if let (Ok(a), Ok(b)) = (resolve_item(item1), resolve_item(item2)) {
-                    if let Some(p) = CanonicalItemUrl::parse(&a).and_then(|c| c.parent()) { parents.insert(p); }
-                    if let Some(p) = CanonicalItemUrl::parse(&b).and_then(|c| c.parent()) { parents.insert(p); }
+                    if let Some(p) = a.parent() { parents.insert(p); }
+                    if let Some(p) = b.parent() { parents.insert(p); }
                 }
             }
         }
@@ -525,7 +521,7 @@ async fn rpc_post(
             .filter_map(|p| {
                 let before = pre_rankings.get(p)?;
                 let after = crate::scope_rank::build_children_rankings(content, p);
-                compute_scope_rank_changes(p.as_str(), before, &after, &room_key)
+                compute_scope_rank_changes(p, before, &after, &room_key)
             })
             .collect();
         if v.is_empty() { None } else { Some(v) }
@@ -638,8 +634,8 @@ async fn rpc_check(
         for s in &v.doc.statements {
             if let dsl::Stmt::Vote { item1, item2, .. } = s {
                 if let (Ok(a), Ok(b)) = (resolve_item(item1), resolve_item(item2)) {
-                    if let Some(p) = CanonicalItemUrl::parse(&a).and_then(|c| c.parent()) { parents.insert(p); }
-                    if let Some(p) = CanonicalItemUrl::parse(&b).and_then(|c| c.parent()) { parents.insert(p); }
+                    if let Some(p) = a.parent() { parents.insert(p); }
+                    if let Some(p) = b.parent() { parents.insert(p); }
                 }
             }
         }
@@ -961,7 +957,7 @@ fn rpc_search(reduced: &ReducerState, q: &str, limit: usize, principal: Option<&
 async fn rpc_get_pair(state: &AppState, room: String, parent_path: String) -> Result<RpcResult, RpcErr> {
     let scope = scope_from_room_wire(&room);
     let reduced_arc = state.reduced.clone();
-    let pool: Vec<String> = {
+    let pool: Vec<CanonicalItemUrl> = {
         let reduced = reduced_arc.read().await;
         let content = content_for_room(&reduced, &room);
         let tmp = if parent_path.trim().is_empty() {
@@ -970,12 +966,11 @@ async fn rpc_get_pair(state: &AppState, room: String, parent_path: String) -> Re
             Some(parent_path.clone())
         };
         let specs = parse_parent_specs(tmp.as_ref());
-        let raw_pool: Vec<CanonicalItemUrl> = if specs.is_empty() {
+        if specs.is_empty() {
             content.ranking_group.idx_to_item.clone()
         } else {
             crate::scope_rank::resolve_scope(content, &specs)
-        };
-        raw_pool.into_iter().map(|it| it.0).collect()
+        }
     };
     if pool.len() < 2 {
         return Err((
@@ -983,31 +978,30 @@ async fn rpc_get_pair(state: &AppState, room: String, parent_path: String) -> Re
             Some("add items via ingest".into()),
         ));
     }
-    let selected: Option<(String, String)> = {
+    let selected: Option<(CanonicalItemUrl, CanonicalItemUrl)> = {
         let mut reduced = reduced_arc.write().await;
         let content = reduced.content.entry(scope.clone()).or_default();
         let group = &mut content.ranking_group;
         if group.idx_to_item.is_empty() {
-            pick_random_distinct(&pool)
+            pick_random_distinct_canonical(&pool)
         } else {
             let mut rng = rand::thread_rng();
-            let idxs: Vec<usize> = pool.iter()
-                .filter_map(|it| {
-                    let key = CanonicalItemUrl(it.clone());
-                    group.item_to_idx.get(&key).copied()
-                })
+            let idxs: Vec<usize> = pool
+                .iter()
+                .filter_map(|it| group.item_to_idx.get(it).copied())
                 .collect();
             let ranked = ranked_items_subset(group, &idxs, 10000, 1e-8);
-            let ranked_set: HashSet<String> = ranked.iter().map(|r| r.item.as_str().to_string()).collect();
-            let unsorted: Vec<String> = pool.iter()
+            let ranked_set: HashSet<CanonicalItemUrl> = ranked.iter().map(|r| r.item.clone()).collect();
+            let unsorted: Vec<CanonicalItemUrl> = pool
+                .iter()
                 .filter(|it| !ranked_set.contains(*it))
                 .cloned()
                 .collect();
-            let mut pick: Option<(String, String)> = None;
+            let mut pick: Option<(CanonicalItemUrl, CanonicalItemUrl)> = None;
             if !unsorted.is_empty() {
                 if let Some(left) = unsorted.choose(&mut rng).cloned() {
-                    let mut candidates: Vec<String> = if !ranked.is_empty() {
-                        ranked.iter().map(|r| r.item.as_str().to_string()).collect()
+                    let mut candidates: Vec<CanonicalItemUrl> = if !ranked.is_empty() {
+                        ranked.iter().map(|r| r.item.clone()).collect()
                     } else {
                         pool.clone()
                     };
@@ -1021,21 +1015,21 @@ async fn rpc_get_pair(state: &AppState, room: String, parent_path: String) -> Re
                     let a = ranked[i].item.as_str();
                     let b = ranked[i + 1].item.as_str();
                     if a != b && !is_pair_voted(group, a, b) {
-                        pick = Some((a.to_string(), b.to_string()));
+                        pick = Some((ranked[i].item.clone(), ranked[i + 1].item.clone()));
                         break;
                     }
                 }
                 if pick.is_none() {
                     for _ in 0..64 {
                         let (Some(a), Some(b)) = (pool.choose(&mut rng).cloned(), pool.choose(&mut rng).cloned()) else { break; };
-                        if a != b && !is_pair_voted(group, &a, &b) {
+                        if a != b && !is_pair_voted(group, a.as_str(), b.as_str()) {
                             pick = Some((a, b));
                             break;
                         }
                     }
                 }
             }
-            pick.or_else(|| pick_random_distinct(&pool))
+            pick.or_else(|| pick_random_distinct_canonical(&pool))
         }
     };
     let Some((left, right)) = selected else {
@@ -1043,8 +1037,8 @@ async fn rpc_get_pair(state: &AppState, room: String, parent_path: String) -> Re
     };
     let reduced = reduced_arc.read().await;
     let content = content_for_room(&reduced, &room);
-    let left_key = CanonicalItemUrl(left.clone());
-    let right_key = CanonicalItemUrl(right.clone());
+    let left_key = left.clone();
+    let right_key = right.clone();
     let lb = content.item_bodies.get(&left_key).cloned();
     let rb = content.item_bodies.get(&right_key).cloned();
     let th: Vec<String> = content
@@ -1058,8 +1052,8 @@ async fn rpc_get_pair(state: &AppState, room: String, parent_path: String) -> Re
         .collect();
     let cs = compute_connectivity_stats(&content.ranking_group, &pool);
     Ok(RpcResult::Pair(PairResponse {
-        left: GardenItemUrl::from_storage_str(&left, &room),
-        right: GardenItemUrl::from_storage_str(&right, &room),
+        left: GardenItemUrl::from_stored(&left, &room),
+        right: GardenItemUrl::from_stored(&right, &room),
         left_body: lb,
         right_body: rb,
         threads: th,
@@ -1554,7 +1548,7 @@ pub async fn handle_rpc_batch(
                     for r in items {
                         let pct = want_percent.then(|| ((r.score - bot) / range * 100.0).clamp(0.0, 100.0));
                         ranked.push(RankRow {
-                            item: GardenItemUrl::from_storage_str(r.item.as_str(), &room),
+                            item: GardenItemUrl::from_stored(&r.item, &room),
                             score: r.score,
                             percent: pct,
                         });
@@ -1562,11 +1556,11 @@ pub async fn handle_rpc_batch(
                 }
 
                 let ranked_total = ranked.len();
-                let mut unranked: Vec<String> = content
+                let mut unranked: Vec<CanonicalItemUrl> = content
                     .items
                     .iter()
                     .filter(|it| !group.item_to_idx.contains_key(*it))
-                    .map(|it| it.as_str().to_string())
+                    .cloned()
                     .collect();
                 unranked.sort();
                 let unranked_total = unranked.len();
@@ -1574,7 +1568,7 @@ pub async fn handle_rpc_batch(
                 let page: Vec<RankRow> = ranked
                     .into_iter()
                     .chain(unranked.into_iter().map(|it| RankRow {
-                        item: GardenItemUrl::from_storage_str(&it, &room),
+                        item: GardenItemUrl::from_stored(&it, &room),
                         score: 0.0,
                         percent: want_percent.then_some(0.0),
                     }))
@@ -1722,7 +1716,7 @@ pub async fn handle_rpc_batch(
                     .filter(|p| !parents.contains(p.as_str()))
                     .map(|p| GardenItemUrl::from_stored(p, &room))
                     .collect();
-                paths.sort_by(|a, b| a.as_str().cmp(b.as_str()));
+                paths.sort();
                 line_ok(RpcResult::Leaves(LeavesResponse { paths }))
                 }
             },
diff --git a/server/src/api/validate.rs b/server/src/api/validate.rs
index a51c783ee9785569b5a44c0b1572471fe00d174b..3c7aa80fd485cf247231fe5c5c7e6fea62f22534 100644
--- a/server/src/api/validate.rs
+++ b/server/src/api/validate.rs
@@ -52,7 +52,7 @@ pub fn validate_ingest_document(
     };
 
     let ts = super::helpers::now_ms();
-    let mut defined_in_doc: HashSet<String> = HashSet::new();
+    let mut defined_in_doc: HashSet<CanonicalItemUrl> = HashSet::new();
 
     for s in &doc.statements {
         match s {
@@ -66,14 +66,14 @@ pub fn validate_ingest_document(
                 let Some(body_text) = body else {
                     return Err((
                         StatusCode::BAD_REQUEST,
-                        format!("item missing body: {}", GardenItemUrl::from_storage_str(&item, room_wire)),
+                        format!("item missing body: {}", GardenItemUrl::from_stored(&item, room_wire)),
                         Some("items must be declared with bodies, e.g. `~/path/item { ... }`".to_string()),
                     ));
                 };
                 if body_text.trim().is_empty() {
                     return Err((
                         StatusCode::BAD_REQUEST,
-                        format!("item body is empty: {}", GardenItemUrl::from_storage_str(&item, room_wire)),
+                        format!("item body is empty: {}", GardenItemUrl::from_stored(&item, room_wire)),
                         Some("write at least one sentence inside `{ ... }`".to_string()),
                     ));
                 }
@@ -102,11 +102,8 @@ pub fn validate_ingest_document(
                 };
                 let missing: Vec<String> = [&a, &b]
                     .into_iter()
-                    .filter(|it| {
-                        let key = CanonicalItemUrl((*it).clone());
-                        !defined_in_doc.contains(*it) && !item_exists(&key)
-                    })
-                    .map(|it| GardenItemUrl::from_storage_str(it, room_wire).into_inner())
+                    .filter(|it| !defined_in_doc.contains(*it) && !item_exists(it))
+                    .map(|it| GardenItemUrl::from_stored(it, room_wire).into_inner())
                     .collect();
                 if !missing.is_empty() {
                     return Err((
@@ -120,11 +117,8 @@ pub fn validate_ingest_document(
                 }
                 let missing_body: Vec<String> = [&a, &b]
                     .into_iter()
-                    .filter(|it| {
-                        let key = CanonicalItemUrl((*it).clone());
-                        !defined_in_doc.contains(*it) && !body_exists(&key)
-                    })
-                    .map(|it| GardenItemUrl::from_storage_str(it, room_wire).into_inner())
+                    .filter(|it| !defined_in_doc.contains(*it) && !body_exists(it))
+                    .map(|it| GardenItemUrl::from_stored(it, room_wire).into_inner())
                     .collect();
                 if !missing_body.is_empty() {
                     return Err((
diff --git a/server/src/html/editor.rs b/server/src/html/editor.rs
index ecdd226b1b17d5f11d1b79f99add5759ef918281..26d648d7c8d6e5d53ce5ed7a051b44856fca2fcf 100644
--- a/server/src/html/editor.rs
+++ b/server/src/html/editor.rs
@@ -130,8 +130,8 @@ pub async fn editor_check(
                 for s in &v.doc.statements {
                     if let crate::dsl::Stmt::Vote { item1, item2, .. } = s {
                         if let (Ok(a), Ok(b)) = (resolve_item(item1), resolve_item(item2)) {
-                            if let Some(p) = crate::path_types::CanonicalItemUrl::parse(&a).and_then(|c| c.parent()) { parents.insert(p); }
-                            if let Some(p) = crate::path_types::CanonicalItemUrl::parse(&b).and_then(|c| c.parent()) { parents.insert(p); }
+                            if let Some(p) = a.parent() { parents.insert(p); }
+                            if let Some(p) = b.parent() { parents.insert(p); }
                         }
                     }
                 }
diff --git a/types/src/paths.rs b/types/src/paths.rs
index 2950a7502255927583fbacdfd2adb700f0b0c221..839684dfa0ecdd6c2572e2d8132ecee4d5c41bf8 100644
--- a/types/src/paths.rs
+++ b/types/src/paths.rs
@@ -3,6 +3,7 @@
 
 use std::borrow::Borrow;
 use std::fmt;
+use std::ops::Deref;
 
 use serde::{Deserialize, Serialize};
 
@@ -319,6 +320,14 @@ impl fmt::Display for GardenItemUrl {
     }
 }
 
+impl Deref for GardenItemUrl {
+    type Target = str;
+
+    fn deref(&self) -> &Self::Target {
+        &self.0
+    }
+}
+
 fn garden_href_string(item: &str, room_wire: &str) -> String {
     let room = room_wire.trim();
     if room.is_empty() || room == "public" {
@@ -386,6 +395,14 @@ impl fmt::Display for ForumThreadUrl {
     }
 }
 
+impl Deref for ForumThreadUrl {
+    type Target = str;
+
+    fn deref(&self) -> &Self::Target {
+        &self.0
+    }
+}
+
 /// `~/a/b` style path for list UIs (paths index `path` field).
 #[derive(Debug, Clone, PartialEq, Eq, Hash, Serialize, Deserialize)]
 #[serde(transparent)]
@@ -412,6 +429,14 @@ impl fmt::Display for TildeOntologyPath {
     }
 }
 
+impl Deref for TildeOntologyPath {
+    type Target = str;
+
+    fn deref(&self) -> &Self::Target {
+        &self.0
+    }
+}
+
 #[cfg(test)]
 mod tests {
     use super::*;
@@ -452,6 +477,13 @@ mod tests {
         assert_eq!(c.tilde_segments(), Vec::<&str>::new());
     }
 
+    #[test]
+    fn garden_item_url_deref_to_str() {
+        let g = GardenItemUrl::from_storage_str("https://slug.social/~/x", "public");
+        let s: &str = &*g;
+        assert_eq!(s, "https://slug.social/~/x");
+    }
+
     #[test]
     fn garden_public_passthrough_https() {
         let u = "https://slug.social/~/a/b";


Side B — contributor: tommy-mor
Side B — commit message:
[bda5f8aa] Remove legacy evidence projection; Evidence envelopes only.

Epoch and commit pages no longer invent history from bare GitDiscovery rows. Production ledger will be wiped to re-emit under the current schema.

Co-authored-by: Cursor <cursoragent@cursor.com>

Side B — unified diff (full patch):
diff --git a/constitution.py b/constitution.py
index 4dd5b9dfbba231d46289c490f91b1dd5b1018bcf..26ba130e885e0e69fb7874ca5c3f07f42100a150 100644
--- a/constitution.py
+++ b/constitution.py
@@ -210,10 +210,10 @@ class Emission:
     distributions: dict   # author -> amount str
     ranking: dict         # author -> score str
     models_used: list
-    discovery_snapshot_id: str = ""  # empty only for pre-discovery ledger history
-    evidence_schema_version: int = 1
-    ranking_run_id: str = ""
-    ranking_event_id: str = ""
+    discovery_snapshot_id: str
+    evidence_schema_version: int
+    ranking_run_id: str
+    ranking_event_id: str
 
 
 @event
@@ -502,26 +502,6 @@ def _epochs_in_ledger() -> list[int]:
     return sorted(epochs)
 
 
-def _legacy_commit_row(commit_id: str) -> tuple[GitDiscovery | None, dict | None]:
-    for discovery in store.read():
-        if not isinstance(discovery, GitDiscovery):
-            continue
-        for commit in discovery.commits:
-            if commit_id_for_oid(commit["oid"]) == commit_id:
-                return discovery, commit
-    return None, None
-
-
-def _legacy_observation(commit_id: str) -> tuple[GitDiscovery | None, dict | None]:
-    for discovery in store.read():
-        if not isinstance(discovery, GitDiscovery):
-            continue
-        for obs in discovery.observations:
-            if commit_id_for_oid(obs["oid"]) == commit_id:
-                return discovery, obs
-    return None, None
-
-
 def build_pairwise_prompt(side_a: dict, side_b: dict) -> str:
     return f"""You are ranking contributions to an open source project.
 Compare these two sides (each may be one or more commits). Decide which side contributed more.
@@ -2040,7 +2020,7 @@ def _strip_heavy_fields(obj: dict) -> dict:
 
 @app.get("/api/ledger")
 async def get_ledger(offset: int = 0, limit: int = 100, full: int = 0):
-    """List of ledger dicts (backward-compatible). Heavy blobs stripped unless full=1."""
+    """List of ledger dicts. Heavy blobs stripped unless full=1."""
     limit = max(1, min(limit, 500))
     rows = []
     for e in store.read()[offset:offset + limit]:
@@ -2274,23 +2254,25 @@ async def epochs_index():
     epochs = _epochs_in_ledger()
     rows = []
     for epoch in epochs:
-        discovery = _discovery_for_epoch(epoch)
         emission = _emission_for_epoch(epoch)
-        evidence_n = sum(
-            1 for e in evidence_by_kind() if e.epoch == epoch
+        evidence_n = sum(1 for e in evidence_by_kind() if e.epoch == epoch)
+        disc = next(
+            (
+                e for e in evidence_by_kind("git.discovery_completed")
+                if e.epoch == epoch
+            ),
+            None,
         )
         detail = []
-        if discovery:
+        if disc:
             detail.append(
-                f"{len(discovery.commits)} eligible / "
-                f"{len(discovery.observations)} observed"
+                f"{disc.payload.get('eligible_count', 0)} eligible / "
+                f"{disc.payload.get('observation_count', 0)} observed"
             )
         if emission:
             detail.append(f"emitted {emission.total_emitted}")
         if evidence_n:
             detail.append(f"{evidence_n} evidence events")
-        elif discovery or emission:
-            detail.append("legacy (no Evidence envelopes)")
         rows.append(["li",
             _a(_evidence_path("epoch", str(epoch)), f"epoch {epoch}"),
             " — ",
@@ -2307,37 +2289,37 @@ async def epochs_index():
 
 @app.get("/epochs/{epoch}")
 async def epoch_detail(epoch: int):
-    discovery = _discovery_for_epoch(epoch)
     emission = _emission_for_epoch(epoch)
     evidence_rows = [e for e in evidence_by_kind() if e.epoch == epoch]
+    if not evidence_rows and emission is None:
+        return _evidence_page(f"epoch {epoch}", [
+            _evidence_nav(),
+            ["h1", f"epoch {epoch}"],
+            ["p.note", "No evidence for this epoch."],
+        ])
+
     commit_evs = [e for e in evidence_rows if e.kind == "git.commit"]
     comparison_evs = [e for e in evidence_rows if e.kind == "comparison.input"]
     judgment_evs = [e for e in evidence_rows if e.kind == "llm.judgment"]
+    discovery_ev = next(
+        (e for e in evidence_rows if e.kind == "git.discovery_completed"), None
+    )
     ranking_started = next(
         (e for e in evidence_rows if e.kind == "ranking.started"), None
     )
     ranking_completed = next(
         (e for e in evidence_rows if e.kind == "ranking.completed"), None
     )
-    legacy = not evidence_rows and (discovery is not None or emission is not None)
-
-    commit_links: list[tuple[str, str]] = []
-    if commit_evs:
-        for e in commit_evs:
-            cid = e.payload.get("commit_id") or ""
-            label = (
-                f"{e.payload.get('oid', cid)[:24]} "
-                f"({e.payload.get('contributor', '?')})"
-            )
-            commit_links.append((label, _evidence_path("commit", cid)))
-    elif discovery:
-        for c in discovery.commits:
-            cid = commit_id_for_oid(c["oid"])
-            commit_links.append((
-                f"{c['oid'][:24]} ({c.get('contributor', '?')})",
-                _evidence_path("commit", cid),
-            ))
 
+    commit_links = [
+        (
+            f"{e.payload.get('oid', e.payload.get('commit_id', ''))[:24]} "
+            f"({e.payload.get('contributor', '?')})",
+            _evidence_path("commit", e.payload["commit_id"]),
+        )
+        for e in commit_evs
+        if e.payload.get("commit_id")
+    ]
     comparison_links = [
         (
             e.payload.get("summary") or e.payload.get("comparison_id", e.event_id),
@@ -2360,16 +2342,13 @@ async def epoch_detail(epoch: int):
     ]
 
     excluded = []
-    if discovery:
-        for obs in discovery.observations:
+    if discovery_ev:
+        for obs in discovery_ev.payload.get("observations") or []:
             if obs.get("eligible"):
                 continue
             oid = obs.get("oid", "?")
             reason = obs.get("exclusion_reason") or "excluded"
-            excluded.append(["li",
-                f"{oid[:28]} — {reason} — ",
-                ["span.note", "legacy evidence unavailable"],
-            ])
+            excluded.append(["li", f"{oid[:28]} — {reason}"])
 
     ranking_nodes: list = []
     if ranking_completed:
@@ -2388,21 +2367,10 @@ async def epoch_detail(epoch: int):
                 indent=2, sort_keys=True,
             )],
         ]
-    elif emission:
-        if legacy and len(emission.ranking or {}) <= 1:
-            ranking_nodes.append(["p.note",
-                "Single-contributor epoch — no LLM judgments."
-            ])
-        ranking_nodes.extend([
-            ["p", "Projected from Emission (no ranking Evidence event)."],
-            ["pre.blob", json.dumps(emission.ranking, indent=2, sort_keys=True)],
-        ])
+    elif ranking_started:
+        ranking_nodes = [["p.note", f"Ranking started: {ranking_started.event_id}"]]
     else:
-        ranking_nodes = [["p.note", "No ranking recorded."]]
-    if ranking_started and not ranking_completed:
-        ranking_nodes.insert(0, ["p.note",
-            f"Ranking started: {ranking_started.event_id}"
-        ])
+        ranking_nodes = [["p.note", "No ranking evidence."]]
 
     if emission:
         emission_node = _dl_rows([
@@ -2411,44 +2379,41 @@ async def epoch_detail(epoch: int):
             ("pool_after", emission.pool_after),
             ("discovery_snapshot_id", emission.discovery_snapshot_id),
             ("ranking_run_id", emission.ranking_run_id or None),
+            ("ranking_event_id", emission.ranking_event_id or None),
             ("models_used", ", ".join(emission.models_used or [])),
             ("distributions", json.dumps(emission.distributions, sort_keys=True)),
         ])
     else:
         emission_node = ["p.note", "No emission for this epoch."]
 
-    single_contributor = False
-    if discovery:
-        single_contributor = len({c.get("contributor") for c in discovery.commits}) <= 1
-    elif emission:
-        single_contributor = len(emission.ranking or {}) <= 1
+    contributors = {
+        e.payload.get("contributor")
+        for e in commit_evs
+        if e.payload.get("contributor")
+    }
+    no_comparisons_note = "No comparisons."
+    if len(contributors) <= 1:
+        no_comparisons_note += " Single-contributor — no LLM judgments."
 
     body = [
         _evidence_nav(),
         ["div.eyebrow", f"epoch {epoch}"],
         ["h1", f"epoch {epoch}"],
     ]
-    if legacy:
-        body.append(["p.note",
-            "Legacy epoch: projected from GitDiscovery/Emission without Evidence "
-            "envelopes. Eligible commits use discovery patches; discarded observation "
-            "metadata is marked legacy evidence unavailable. Single-contributor "
-            "epochs have no LLM judgments."
-        ])
-    if discovery:
+    if discovery_ev:
         body.extend([
             ["h2", "discovery"],
             _dl_rows([
-                ("snapshot_id", discovery.snapshot_id),
-                ("config_digest", discovery.config_digest),
-                ("initial_snapshot", discovery.initial_snapshot),
-                ("observations", len(discovery.observations)),
-                ("eligible", len(discovery.commits)),
+                ("snapshot_id", discovery_ev.payload.get("snapshot_id")),
+                ("config_digest", discovery_ev.payload.get("config_digest")),
+                ("observations", discovery_ev.payload.get("observation_count")),
+                ("eligible", discovery_ev.payload.get("eligible_count")),
+                ("event", _a(
+                    _evidence_path("event", discovery_ev.event_id),
+                    discovery_ev.event_id,
+                )),
             ]),
         ])
-    no_comparisons_note = "No comparisons."
-    if single_contributor:
-        no_comparisons_note += " Single-contributor — no LLM judgments."
     body.extend([
         ["h2", "commits"],
         _link_list(commit_links),
@@ -2471,92 +2436,42 @@ async def epoch_detail(epoch: int):
 @app.get("/commits/{commit_id}")
 async def commit_detail(commit_id: str):
     ev = find_evidence_payload("git.commit", "commit_id", commit_id)
-    discovery, legacy_row = (None, None)
     if not ev:
-        discovery, legacy_row = _legacy_commit_row(commit_id)
-    if not ev and not legacy_row:
-        discovery, obs = _legacy_observation(commit_id)
-        if obs is not None:
-            epoch = discovery.epoch if discovery else "?"
-            return _evidence_page(f"commit {commit_id[:24]}", [
-                _evidence_nav(
-                    _a(_evidence_path("epoch", str(epoch)), f"epoch {epoch}")
-                ),
-                ["div.eyebrow", "commit"],
-                ["h1", commit_id],
-                ["p.note", "legacy evidence unavailable"],
-                _dl_rows([
-                    ("oid", obs.get("oid")),
-                    ("eligible", obs.get("eligible")),
-                    ("exclusion_reason", obs.get("exclusion_reason")),
-                    ("epoch", str(epoch)),
-                ]),
-            ])
         return _evidence_page("commit not found", [
             _evidence_nav(),
             ["h1", "commit not found"],
             ["p", commit_id],
         ])
-    if ev:
-        p = ev.payload
-        epoch = ev.epoch
-        oid = p.get("oid", "")
-        contributor = p.get("contributor", "")
-        message = _blob_text(p.get("message"))
-        patch = _blob_text(p.get("patch"))
-        meta = _dl_rows([
+    p = ev.payload
+    epoch = ev.epoch
+    return _evidence_page(f"commit {commit_id[:24]}", [
+        _evidence_nav(_a(_evidence_path("epoch", str(epoch)), f"epoch {epoch}")),
+        ["div.eyebrow", "commit"],
+        ["h1", commit_id],
+        _dl_rows([
             ("commit_id", commit_id),
-            ("oid", oid),
-            ("contributor", contributor),
+            ("oid", p.get("oid", "")),
+            ("contributor", p.get("contributor", "")),
             ("patch_sha256", p.get("patch_sha256")),
             ("patch_identity", p.get("patch_identity")),
             ("committer_timestamp_ms", p.get("committer_timestamp_ms")),
             ("evidence_event", _a(_evidence_path("event", ev.event_id), ev.event_id)),
             ("epoch", _a(_evidence_path("epoch", str(epoch)), str(epoch))),
-        ])
-        legacy_note = None
-    else:
-        assert legacy_row is not None and discovery is not None
-        epoch = discovery.epoch
-        oid = legacy_row.get("oid", "")
-        contributor = legacy_row.get("contributor", "")
-        message = legacy_row.get("message") or ""
-        patch = legacy_row.get("patch") or ""
-        meta = _dl_rows([
-            ("commit_id", commit_id),
-            ("oid", oid),
-            ("contributor", contributor),
-            ("patch_sha256", legacy_row.get("patch_sha256")),
-            ("patch_identity", legacy_row.get("patch_identity")),
-            ("committer_timestamp_ms", legacy_row.get("committer_timestamp_ms")),
-            ("epoch", _a(_evidence_path("epoch", str(epoch)), str(epoch))),
-            ("source", "legacy GitDiscovery projection"),
-        ])
-        legacy_note = ["p.note", "Projected from GitDiscovery — Evidence envelope absent."]
-    return _evidence_page(f"commit {commit_id[:24]}", [
-        _evidence_nav(_a(_evidence_path("epoch", str(epoch)), f"epoch {epoch}")),
-        ["div.eyebrow", "commit"],
-        ["h1", commit_id],
-        *([legacy_note] if legacy_note else []),
-        meta,
+        ]),
         ["p", _a(f"/commits/{commit_id}/patch", "download patch")],
         ["h2", "message"],
-        _pre_blob(message),
+        _pre_blob(_blob_text(p.get("message"))),
         ["h2", "patch"],
-        _pre_blob(patch),
+        _pre_blob(_blob_text(p.get("patch"))),
     ])
 
 
 @app.get("/commits/{commit_id}/patch")
 async def commit_patch_download(commit_id: str):
     ev = find_evidence_payload("git.commit", "commit_id", commit_id)
-    if ev:
-        raw = _decode_blob(ev.payload.get("patch")) or _blob_text(ev.payload.get("patch")).encode()
-    else:
-        _, row = _legacy_commit_row(commit_id)
-        if not row:
-            return PlainTextResponse("not found", status_code=404)
-        raw = (row.get("patch") or "").encode("utf-8")
+    if not ev:
+        return PlainTextResponse("not found", status_code=404)
+    raw = _decode_blob(ev.payload.get("patch")) or _blob_text(ev.payload.get("patch")).encode()
     return Response(
         content=raw,
         media_type="text/plain; charset=utf-8",
@@ -3226,7 +3141,9 @@ async def watch():
                 ["div.controls",
                     ["button#play", {"type": "button", "disabled": "true"}, "▶ Play live feed"],
                     ["button#pause", {"type": "button"}, "Ⅱ Pause feed"],
-                    ["span.note", "Pause stops local updates; the constitutional process continues."],
+                    ["span.note",
+                     "Play/pause only affect this browser feed — they do not run or stop emission. "
+                     "The server emits at each epoch boundary; status above is that process."],
                 ],
             ],
             ["section.panel",
diff --git a/tests/integration.clj b/tests/integration.clj
index 204b0fedfbad57846205cb140c4ed9406b5f8e0e..11140e95a8cd86768a661a30e6b29a3ddfc95fcc 100644
--- a/tests/integration.clj
+++ b/tests/integration.clj
@@ -247,7 +247,10 @@
                :decay_rate   "0.003253356063468"
                :distributions {"alice" "381.4282558872" "bob" "190.7141279436"}
                :ranking       {"alice" "0.6666" "bob" "0.3334"}
-               :models_used   ["mock/chat-v1" "mock/chat-v2"]}]
+               :models_used   ["mock/chat-v1" "mock/chat-v2"]
+               :evidence_schema_version 2
+               :ranking_run_id "seed-rank"
+               :ranking_event_id "seed-rank-ev"}]
     (spit path (str (json/generate-string entry) "\n"))))
 
 ;; ---------------------------------------------------------------------------
diff --git a/tests/test_evidence.py b/tests/test_evidence.py
index 76adcd343eec3a65dec58fda93b068356fb1adcd..7cb74fd7fb1bcd0ab17d2cbe2e6ca7e61e164221 100644
--- a/tests/test_evidence.py
+++ b/tests/test_evidence.py
@@ -1,4 +1,4 @@
-"""HTML evidence graph: byte fidelity, resume, legacy projection, private ack."""
+"""HTML evidence graph: byte fidelity, resume, and Evidence-only pages."""
 
 from __future__ import annotations
 
@@ -86,55 +86,10 @@ def test_html_evidence_pages_escape_and_link(evidence_store):
     assert patch_resp.headers["content-disposition"].startswith("attachment")
 
 
-def test_legacy_epoch_projects_from_gitdiscovery_without_evidence(evidence_store):
-    oid = "sha1:" + "b" * 40
-    discovery = c.GitDiscovery(
-        schema_version=1,
-        epoch=3,
-        snapshot_id="legacy-snap",
-        timestamp_ms=1,
-        config_digest="d",
-        initial_snapshot=True,
-        configuration={"repositories": [], "contributors": {}},
-        repositories=[],
-        observations=[{
-            "oid": oid,
-            "eligible": False,
-            "exclusion_reason": "before_genesis",
-            "first_sources": [],
-        }, {
-            "oid": "sha1:" + "c" * 40,
-            "eligible": True,
-            "exclusion_reason": None,
-            "first_sources": [],
-        }],
-        commits=[{
-            "oid": "sha1:" + "c" * 40,
-            "contributor": "tommy-mor",
-            "message": "eligible",
-            "patch": "diff --git a/x b/x\n",
-        }],
-    )
-    emission = c.Emission(
-        epoch=3,
-        timestamp_ms=2,
-        pool_before="1",
-        total_emitted="0.1",
-        pool_after="0.9",
-        decay_rate="0.1",
-        distributions={"tommy-mor": "0.1"},
-        ranking={"tommy-mor": "1"},
-        models_used=[],
-        discovery_snapshot_id="legacy-snap",
-    )
-    asyncio.run(c.store.append(discovery))
-    asyncio.run(c.store.append(emission))
-    html = asyncio.run(c.epoch_detail(3)).body.decode()
-    assert "legacy" in html.lower()
-    assert "Single-contributor" in html or "no LLM" in html
-    excluded_id = c.commit_id_for_oid(oid)
-    excluded_html = asyncio.run(c.commit_detail(excluded_id)).body.decode()
-    assert "legacy evidence unavailable" in excluded_html.lower()
+def test_commit_without_evidence_is_not_found(evidence_store):
+    html = asyncio.run(c.commit_detail("c_missing")).body.decode()
+    assert "commit not found" in html
+    assert "legacy" not in html.lower()
 
 
 def test_ranking_resume_skips_duplicate_provider_calls(evidence_store, monkeypatch):
@@ -202,6 +157,10 @@ def test_epochs_index_lists_epochs(evidence_store):
         distributions={},
         ranking={},
         models_used=[],
+        discovery_snapshot_id="snap",
+        evidence_schema_version=c.EVIDENCE_SCHEMA_VERSION,
+        ranking_run_id="r",
+        ranking_event_id="e",
     )))
     html = asyncio.run(c.epochs_index()).body.decode()
     assert "/epochs/3" in html