Storing the names is inefficient.  Just create your own id that cannot change and use it.  For a few rows, won't matter much.  For lots of rows, could be a huge gain.  I don't see a need to distinguish object type, so I don't, I just have a single names table that provides the id for any name.

Although even if I chose to store the name, I would also store the original id, since certain internal / trace / relationship tables use only id not the name.  Compress the row and it'll be no more than 5 bytes the vast majority of the time.