diff --git a/work_note2 b/work_note2
new file mode 100644
index 0000000..ab93836
--- /dev/null
+++ b/work_note2
@@ -0,0 +1,48 @@
+---
+  Still Needs Attention
+  
+  1. float altitude is now listed twice in target_data_structure (lines 130 and 136).
+  float altitude; // in meters          ← line 130, was originally "float height"
+  ...
+  float altitude; // meters, but 0 for boats  ← line 136
+  It looks like the original float height (structural height of the vessel) got accidentally renamed to altitude,
+  creating a duplicate. The structural height of the vessel is a distinct concept from flight altitude. float 
+  height should be restored at line 130, or the duplicate removed if structural height is intentionally dropped
+  from the structure.
+  
+  2. getMaterialProfile parameter still mismatches the switch.
+  The function signature is getMaterialProfile(int aisTypeCode) but the switch operates on vessel_type. In actual
+  C++ code, vessel_type would be out of scope inside the function — the parameter would need to be named
+  vessel_type, or the switch would need to use aisTypeCode. Minor, but would cause a compile error.
+
+  3. Missing semicolons on the last two SSBO struct fields (lines 154–155).
+  int32_t police_boat  // no semicolon
+  int type             // no semicolon
+  Would be compile errors in real code.
+  
+  4. binary may_need_update in the database schema (line 1001) is still not a proper type.
+  int32_t police_boat was correctly fixed, but may_need_update still uses binary. In PostgreSQL this should be
+  BOOLEAN or SMALLINT. Should be consistent with police_boat.
+
+  5. timestamp elapsed_seconds — the name and type contradict each other.
+  A PostgreSQL TIMESTAMP stores a calendar date+time (e.g., 2026-05-28 14:30:00), not a count of elapsed seconds.
+  If the intent is seconds-since-epoch, the column type should be BIGINT (and the name elapsed_seconds makes
+  sense). If the intent is an actual timestamp, use TIMESTAMP and name it last_seen or similar. The current
+  combination is confusing.
+
+  ---
+  Minor
+
+  - Line 1185: typo "shouse" → "should"
+  - The unresolved architectural question from before (does traffic_cop filter for the currently active scope
+  only, or does it maintain separate lists per radar?) is still implicit. This may become clear once the simulator
+   section is written, so it may not need addressing now.
+
+✻ Crunched for 59s
+
+──────────────────────────────────────────────────────────────────────────────────────────────────────────────────
+❯  
+──────────────────────────────────────────────────────────────────────────────────────────────────────────────────
+  ? for shortcuts · ← for agents
+
+