Should we document risks with column names longer than NAMEDATALEN?

[wolfgangwalther]

See PostgREST/postgrest#1937

[steve-chavez]

Hm, perhaps we should document these limitations in a section under "Administration"?

This reminds me of:

• bigint limitation: bigint columns truncated to Javascript limits (52 - 53 bits) postgrest#498
• big json limitation: https://stackoverflow.com/questions/56714274/postgrest-postgresql-cannot-enlarge-string-buffer-message