Where does omero metadata go?

[jluethi]

The omero metadata has been labeled "transitional" in the current spec draft, see https://ngff.openmicroscopy.org/latest/#omero-md.
This means the intention is to remove it in the future, once the actual storage destination for such metadata in OME-NGFF has been figured out (see https://ngff.openmicroscopy.org/latest/#document-conventions). I have no idea what the timeline is for this, it doesn't appear imminent. But we should keep an eye on it and ensure the info we need will be available in the new metadata when it comes up (I strongly suspect this is after v0.5)

Also, we are starting to add some custom metadata, e.g. wavelength_id for channel selection, see #211
We can use this as a conversation topic with the OME-NGFF folks, but I don't think it makes sense to contribute changes to the transitional "omero" spec. Let's just make sure we bring up these additions when the new metadata destination is figured out

[jluethi]

I think our main approach here should be adding experimental key-value pairs in .zattrs to figure out what metadata we want to store in OME-Zarr. Once we have those figured out better, we can contribute it back to the spec.

Some ideas:

• a history attribute that gets an ordered list of what has been applied to an OME-Zarr file with what versions/parameters
• Additional entries as discussed here: Each attribute in metadata needs to exist somewhere else in the OME-Zarr file #212

[jluethi]

Turns out, things move much slower in the spec than I thought in 2022 ;)

Let's open a new issue if there's actually a need to move omero metadata somewhere else