The current spatial convention already acknowledges explicit coordinate arrays through potential future transform types such as "lookup".
This raises a broader design question regarding non-spatial coordinates.
Today, the convention focuses on the two horizontal X/Y dimensions and affine-like spatial referencing. However, many datasets also include other coordinate dimensions such as:
- time
- vertical/depth
- spectral bands
- irregular dimensions
- curvilinear or swath coordinates
The question is therefore more about convention boundaries and composition.
Should non-spatial coordinates be handled:
- directly within the current
spatial convention, potentially requiring a broader scope and maybe a different name,
- or through a separate dedicated convention focused on generic coordinate systems and coordinate arrays?
If the second approach is preferred, it would be useful to discuss:
- what the scope boundary between conventions should be,
- how conventions compose together,
- and what a suitable name for such a generic coordinate convention could be.
Possible naming ideas:
coordinatescoordaxesdimensionsgriddomain
I would be interested in feedback on the intended long-term direction before additional conventions emerge independently.
I would typically expect spatial and multiscales to extend a "coordinate" convention: so the multiscale would be clearer for non-spatial dimensions.
The current
spatialconvention already acknowledges explicit coordinate arrays through potential future transform types such as"lookup".This raises a broader design question regarding non-spatial coordinates.
Today, the convention focuses on the two horizontal X/Y dimensions and affine-like spatial referencing. However, many datasets also include other coordinate dimensions such as:
The question is therefore more about convention boundaries and composition.
Should non-spatial coordinates be handled:
spatialconvention, potentially requiring a broader scope and maybe a different name,If the second approach is preferred, it would be useful to discuss:
Possible naming ideas:
coordinatescoordaxesdimensionsgriddomainI would be interested in feedback on the intended long-term direction before additional conventions emerge independently.
I would typically expect spatial and multiscales to extend a "coordinate" convention: so the multiscale would be clearer for non-spatial dimensions.