EpochPosition: general description

[mcdittmar]

EpochPosition
This object is better described as a consolidation of Measure-s rather than a consolidation of Coordinates.
The individual attributes mostly map to Coordinate level elements, but the structure and intent is to include both the values + errors for each of the component pieces, which are Measure types. See image below from my Malta presentation on the mapping.

EpochPosition.parallax
* This could be associated with a GenericMeasure, but the description is poor (see diagram). I think this needs a more complete description of parallax and its expectations. For example, I think this is assuming a 1AU base. I'll open an issue on Measure to define a Parallax type with proper coordinate space definition and necessary metadata.

[lmichel]

I d not see how to describe it as a consolidation of Measure since the used component are from Coords?

What I would absolutely avoid is to renounce to the flat structure of the class:

• here are the parameters
• here are the correlations
• here are the errors
  I'm aware that's not the way you see it, but this what implementers are waiting for.

[mcdittmar]

I d not see how to describe it as a consolidation of Measure since the used component are from Coords?

I was hoping the diagram would help with that.

Conceptually, EpochPosition + EpochPositionErrors is a flattened view of a collection of Measure-s. About half of the attributes link down to the Coordinates level, the other half (mostly EpochPositionError content, but also pmCosLat_applied ) link back to the Measure model.

A Measure is an ObjectType consisting of a piece holding the coordinate information + a piece holding the associated error information, just as EpochPosition is. Coordinates are all DataTypes.

As for how to describe that.. I'd rephrase the first paragraph from:
"This class (fig4) is a view of Astronomical Coordinates and Coordinate Systems (Rots and Cresitello-Dittmar et al., 2022) components that have been put together to form a consistent description of the position of an object moving over time. It consists of a celestial position, a proper motion, a radial velocity and a parallax"

to

" This class (fig4) is a flattened view of objects/concepts from the Astronomical Measurements Model (Rots and Cresitello-Dittmar et al., 2022) that have been put together to form a consistent description of the position of an object moving over time. It consists of a celestial position, a proper motion, a radial velocity and a parallax and their associated errors encapsulated into the EpochPositionErrors class. The values of these properties are pulled from the underlying Astronomical Coordinates and Coordinate Systems model (Rots and Cresitello-Dittmar et al., 2022)
At a high level the properties map as follows:
- celestial position -> meas:Position
- proper motion -> meas:ProperMotion
- radial velocity -> meas.Velocity
- parallax -> no suitable counterpart at this time
"

The attribute descriptions can be pretty much unchanged, mapping to the corresponding coords attribute.

I will re-iterate that I also think this first paragraph should highlight that the object is designed to support the epoch propagation thread, and those wishing to associate individual properties (eg: Position) should use the PhysicalProperty.

[lmichel]

ok,

Following your suggestion, I reorganized the table of content (by script) in a way that there is a section dedicated to the EP stuff (see on my branch).
I propose to put your text in the head of this section.