You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
A few of us had a quick discussion today relating to asset specific metadata: the ability to specify fields that are Item properties on a per asset basis.
It's discussed here in the context of the proposed "grids" extension: #688 (comment)
but could apply to other fields as well.
We concluded this was a good feature to have, with the following notes:
properties defined at the asset level should always be first defined at the Item level. The asset field, if present, overrides the Item field for that asset
make clear that fields defined per asset will not be searchable....users will search on the Item properties so that should be some sort of "nominal" value
recommendation should include a short list of fields this may make sense for, e.g., grids extension, gsd
otherwise, providers should think carefully on whether or not they should be doing this, or if what they really need is separate Items for this data
This seems mostly complete, although it doesn't clearly state how the relation between asset-level metadata and common metadata + extension fields is. There's a potential conflict between the "title" fields, should be mentioned at least. Also, what's the take on things like eo:bands that differs between properties and asset level? I think that needs some thoughts and explanation.
We probably need to clarify that the asset extension can also inherit the "new" fields.
We just had a use case where we have to specify the "datetime" field on the asset level (i.e. export of a single band time series), which IMHO is reasonable to do although STAC Items are more modeled for data with a single timestamp. And that makes me think of #613 again. ;-)
A few of us had a quick discussion today relating to asset specific metadata: the ability to specify fields that are Item properties on a per asset basis.
It's discussed here in the context of the proposed "grids" extension:
#688 (comment)
but could apply to other fields as well.
We concluded this was a good feature to have, with the following notes:
did I miss anything @m-mohr @cholmes @anayeaye ?
The text was updated successfully, but these errors were encountered: