scipy is considering to mark the netcdf reader in scipy.io as “legacy” (which means “not deprecated, but minimally maintained”). See RFC: declaring netcdf IO functions in `scipy.io` as legacy · Issue #24885 · scipy/scipy · GitHub
I was under the assumption that that was the only reader for netcdf version 3 files, but according to Incremental writes with scipy.io.netcdf_file · Issue #9157 · scipy/scipy · GitHub the difference is mainly performance. In my experience, netcdf version 3 files are mainly read using xarray’s scipy backend,
How much do we care about the change in the maintenance status of the scipy.io netcdf reader?
And if we do care, what should the future of that functionality be?