Review 3

Review 3#

Review authors#

This review was submitted by Jeremy Maitin-Shepard (Google) via email.

Initial feedback#

Regarding the requirement not to use the new metadata with zarr v2 and not to use the old metadata with zarr v3: in Neuroglancer I have already implemented support for the old metadata in zarr v3. That will likely stay. However I would be happy to not extend that support, and just say that if users want support for any new OME metadata they will need to use the new format and zarr v3.

Regarding the version numbering: I think it would help to clarify how readers and writers of the format should handle the version number. For example, if I’m writing OME metadata, should I attempt to determine the oldest version of the metadata that supports all of the functionality I’m using? Should I write multiple redundant versions of the metadata? When reading the metadata, if I encounter a newer version than is known, should I just ignore it altogether or should I attempt to parse it as the latest known version anyway?