Do you agree that adding a field to NIVolume for indicating the original byteorder can be beneficial? Currently after loading from the disk, this information got lost. In Nibabel.py, there is a ni.header.endianness field for this purpose. Since adding such a field to the header definition is tricky in Julia implementation, we can have it in NIVolume struct, and make it default to the current system byteorder, when a new NIVolume is created.
Do you agree that adding a field to NIVolume for indicating the original byteorder can be beneficial? Currently after loading from the disk, this information got lost. In Nibabel.py, there is a ni.header.endianness field for this purpose. Since adding such a field to the header definition is tricky in Julia implementation, we can have it in NIVolume struct, and make it default to the current system byteorder, when a new NIVolume is created.