I have some complaints.

• I don't like the name "H1dot".

• The document author might want the dots to appear in headings only at a certain heading level; in other words, why don't we have a threshold register for this as we do almost all other heading behavior?

• The document author might want dots in the formatted heading but not in cross references to them using groff mm's "new (sic) reference system".  They get unconditionally stored, except at the first level if `H1dot` is cleared.  This is inflexible; it's much easier to not store the dot in the first place than strip it out later.

For example, the package defaults give us outcomes like this.

See chapter 3., page 5.
See chapter 2.1, page 3.

We might consider storing strings for the headings in end-dotted and non-end-dotted forms, trusting the user to (re-)define the Qrf string as needed (and as already encouraged).