notes.md

@@ -2,26 +2,13 @@
 
 ## [6/8/21] UniformSpace refactor
 
-Need to:
-
-1. Identify the currently used methods
-2. Change some names so that the interface makes sense with the Galilean group action (which is not based on the Lie basis directly)
-
-I need a name for the core manifold data: point (but that's used for the point that the user moves), data (allright but generic), manifold point (verbose), action data, manifold state, view state, orientation (not bad, orientation = position + angle which is what it is), configuration
-
-Actually, we can reduce the importance of the notion of a point.
-A point gets meaning only when it is printed.
-Currently, I store a bunch of points on a 2D surface as constrained 3D points.
-So I do save points.
-So I guess points are kind of important.
-I could also save directly the action data for each point, but that would take more space.
-
 So then what are the methods I need:
 
 * `action_data_to_center()` returns the center point associated with the current action data
 * `action_data_to_lie_basis()` returns the lie algebra basis associated with the current action data
+* `action_data_to_tangents(lie basis)` returns the 3 tangents (or 2, since one is null) for each lie algebra. The tangents are in pebble space and belong to the tangent space of the center.
 * `act_on_data(lie vector)` updates
-* `view_pebbles(array of 3D points all lying on a 2D manifold)`
+* `view_pebbles(array of 3D points all lying on the pebble space which is a 2D manifold)`
 * `act_on_pebbles(lie vector)`
 * `killing(lie vector, lie vector)`
 * `lie_bracket(lie vector, lie vector)`
@@ -46,6 +33,20 @@ Should I move to typescript?
 I don't think the class should hold the action data.
 Because it contains more information than just action data, it is more general.
 
+Need to:
+
+1. Identify the currently used methods
+2. Change some names so that the interface makes sense with the Galilean group action (which is not based on the Lie basis directly)
+
+I need a name for the core manifold data: point (but that's used for the point that the user moves), data (allright but generic), manifold point (verbose), action data, manifold state, view state, orientation (not bad, orientation = position + angle which is what it is), configuration
+
+Actually, we can reduce the importance of the notion of a point.
+A point gets meaning only when it is printed.
+Currently, I store a bunch of points on a 2D surface as constrained 3D points.
+So I do save points.
+So I guess points are kind of important.
+I could also save directly the action data for each point, but that would take more space.
+
 
 
 ## [21/5/21] Jekyll tags and categories