No requirements for ESDC v2.1

[dzelge]

• Organise Meeting
• Generate list of requirements

[dzelge]

Requiremnets for ESDC v2.1

• Slightly more precise cube generation for high res cube?
• Cube loads into the ESDL viewer without skew
• EDIT: Allow variables without time coordinate
• Updated chunking and observing the xcube cube standard
• EDIT: xcube optimize teh cubes
• new varibales
• updated variables
• EDIT: Standardise names
• FAIR cubes
  • complete meta data
  • make original meta data json
  • add history
  • add info about cube generation
• EDIT: GLEAM -> update from version 3.2 to 3.3
• EDIT: ERA5 -> update time span back to 1980

[meggart]

Thanks for putting this together. I would add

• make the cube load faster by using consolidated metadata

Updated variables from our side would be

• GLEAM -> update from version 3.2 to 3.3
• ERA5 -> update time span back to 1980

Don't really know what you mean with "Cube without time coordinate" Does this mean removing the time axis from some static variables? I would definitely support this idea. Do you imagine producing a separate dataset holding only the static variables or adding them to the main cube. AFAIK the xarray Dataset model allows including DataArrays sharing some, but not all axes, but I have no idea if this woukld still work with xcube.

[dzelge]

Don't really know what you mean with "Cube without time coordinate" .

Yes, that's what I mean: Allow variables without time dimension. We discussed that during tehj meeting we had in Jena. Some varibales just don't need a time dimension. How we do that is a matter of discussion.

make the cube load faster by using consolidated metadata

This will be part of making the cubes xcube standard. xcube does allow 'non-optimized' cube how we call them but when you create a cubeit will strongly recommend 'optimizing' the cube how we call it. Part of that optimisation is the consolidation of metadata.