One more probably too advanced thing - if your calculation uses inequality
comparisons a lot, you can just sort `str[]`, clear the table, and re-insert
them all to get the property that `string1 < string2` if and only if `wordNum1
< wordNum2`. This of course invalidates any prior wordNum's. So, it is likely
only worth doing if you have multiple "phases" to your calcs and past the first
phase cares about order. That dovetails into [order preserving minimal perfect
hash
functions](https://en.wikipedia.org/wiki/Perfect_hash_function#Order_preservation)
which then connects to information retrieval techniques indexers like Google
use (covered fairly accessibly in Witten et al. 1999 _Managing Gigabytes_ ). {
You probably needn't concern yourself with minimal perfection, though, esp.
since Nim Table uses simple linear probing for collision resolution and has a
strong-ish string hash. } Anyway, "interning" and "corpus representation" are
along a continuum.
