Should John Smith, MD - PhD split its suffixes without a declared delimiter, and can suffix_delimiter be deprecated? #889

Background. suffix_delimiter (#156, 1.x) and Policy.extra_suffix_delimiters (2.0) let a caller declare a separator between post-nominals, such as RN - CRNA. The v1 setting is the HumanName keyword and the Constants attribute, and the facade layers it onto the Policy field. Since then the default has learned most of what the knob was for: the comma tail's credential run, joined credentials (#887, MD/PhD, MD-PhD), and ; as a default delimiter after the first comma (#871).

What the default does today, measured on master, compared with a declared delimiter:

So the fields come out right by default, which was #156's request. The knob still does two things:

• it splits the suffix into separate entries (MD, PhD rather than MD - PhD);
• it drops a stray separator when something other than a credential follows (MD - → MD).

Questions

1. Should -, /, | and & be default suffix delimiters after the first comma, the way Read a semicolon between post-nominals as a comma (D15) #871 made ; one? That gives the default both of the knob's remaining effects.
   • Measure as Read a semicolon between post-nominals as a comma (D15) #871 did: the committed snapshot, the agreement records, PSC, and the local corpora, with ctgov's after-comma text the main population. Use --base.
   • The known risk is a spaced hyphenated surname (Smith - Jones, Should a lone dash be a separator, so John Smith - Jr. reads family 'Smith' rather than '-'? #820). Keeping the defaults to the part after the first comma should keep such a surname out of reach, but the sweep has to show that: the four comma shapes, three orders, and each spacing of the separator (standing alone, or glued on either side).
   • & also joins names (John & Mary Smith), and P3 reads it as a connective. Check that it stays one wherever no credential stands on both sides.
2. If question 1 lands, should suffix_delimiter and extra_suffix_delimiters be deprecated in 2.7, with a DeprecationWarning naming what to use instead, and removed in 3.0? That would add an entry to decisions/3-0-reevaluations. Removing them in a minor release would break callers who set them, so deprecation comes first.
3. Is there a separator a caller would still need to declare once these are defaults? If so, the knob stays and only its docs change.

Not in scope: MD;; PhD (a doubled separator), the full-width ；, and spaced hyphenated surnames as data cleanup (#820).