Add ability to construct a normalized representation of an entire OLD #48
Description
Problem: It is currently difficult to compare two OLD data sets. We need a way to create a normalized representation of an entire OLD as data. Hopefully we can make EDN the canonical format of this data. However, JSON or some other format may prove more practical. If two databases are identical, this normalized and canonical representation should allow us to quickly determine that fact. A normalized dump of two identical OLDs should result in the same bytes on disk. Considerations of Unicode normalization, data history, and sort ordering will come into play here. There may be a nuanced definition of "same" for two OLDs. That is ok.