Streamlining definitions of specs versus the actual data

I am currently going through the code, trying to make the API more consistent.

One thing I noticed is a asymmetry regarding the 'name' of an object.

For attributes and datasets, the name string is part of the *_specs_t struct, while for groups it is part of the escdf_group_t struct. Was there any particular reason for that choice, of shall I move it into the escdf_group_specs_t struct?