Standardize naming of groups and nodes in output files

So far, we have named the groups and nodes of our output files slightly randomly. I propose that we name them in a standard manner. We can discuss other options, but my preferred one is /data_type/node_name, where

• data_type can be RWF, PMAPS, KDST, HDST, DDST, etc, and
• node_name will be something descriptive of what it holds (like S1Pmt for pmaps) or simply data for when there is a single table, and there is no obvious better name for the node.

Of course, this will need to update some of our reader functions so they are backward-compatible, but it might be worth the effort. Thoughts?