"image-label" not camel case

[constantinpape]

Most of the names in the spec are ~~snake case~~ camel case. However, image-label is not, see https://ngff.openmicroscopy.org/latest/#label-md. Is there a particular reason for this? If not, could it still be changed? I find it fairly confusing. (But I understand that you might not want to do it for legacy reasons; I don't feel super strongly about it.)

Edit: the same is true for label-value.

[constantinpape]

Slightly related minor point: The text in https://ngff.openmicroscopy.org/latest/#label-md does not mention version, but the example json contains it.

[sbesson]

My assumption is that it's historical and according to https://github.com/ome/ngff/pull/70, if there was a change it should rather be imageLabel , imageValue .

This raises the question of how we want to review the keys introduced before the naming convention. Either supporting the legacy name and the new name for a while or making a clear breaking change both have merits

[constantinpape]

My assumption is that it's historical and according to #70, if there was a change it should rather be imageLabel , imageValue .

You're right, we decided to go for camelCase. So indeed it should be imageLabel and imageValue.

This raises the question of how we want to review the keys introduced before the naming convention. Either supporting the legacy name and the new name for a while or making a clear breaking change both have merits

My vote would be for a clear breaking change (which of course would mean a minor version bump). But again, I don't have very strong opinions on this since I'm not the one who'd need to change the code for this ;).