Derived Data Traceability

[cholmes]

One key thing to get right in cloud native geospatial is how to trace back to a 'source' image file. That could be an ortho-ed image back to its level 1 unrectified source, for those who want to apply their own DEM. It could be an NDVI analysis that derives from a source image, so users can trace back to the original. It could even be a mosaic of images, to trace back where it came from.

Doing this fully would need to be linked to metadata about the processes that are run on the data, linking a derived data set to both the source data and the parameters / process that created it.

But it might be good to add a 'source' field that can link back to where the data came from. If the image is the most raw source then it could express that in the field. It would also be good to think through caching on multiple clouds, if we use the 'source' field or another field to indicate that the image is the same.