Should john v read V as the family name, as JOHN V and john x do? #721

In a name written wholly in lower case, a final v is still read as a generational suffix even when no word is left for the family name:

>>> parse("john v")    # given 'john', suffix 'v', family ''   (reports given-or-family)
>>> parse("JOHN V")    # given 'JOHN', family 'V'               (reports suffix-or-name)
>>> parse("john x")    # given 'john', family 'x'               (reports suffix-or-name)
>>> parse("john i")    # given 'john', family 'i'               (reports conjunction-or-initial)

rules/S10 (the trailing roman numeral, #718/#719) says a one-letter numeral needs a word to spare, and that a name written in one case says nothing about any letter in it. john v is the one spelling that skips the count. Lowercase v is a listed suffix word and isn't initial-shaped, so the trailing walk takes it as an ordinary suffix word before the numeral fork is reached. i reaches the fork because a one-case i is tagged as an initial (it's also P3's connective). x reaches it because it's in no vocabulary.

The same cause means john smith v reports nothing, while JOHN SMITH V reports suffix-or-name.

Proposed: in a one-case name, send a one-letter numeral to the numeral fork before the suffix-word test, so john v reads family v and john smith v reports. This would retire S10's first Accepted clause. The population that moves is still to be measured.