Repository navigation
Should John Smith (;.MD) read suffix ;.MD? (an & or ; beside a period credential) #1172
Description
_period_chunks admits an & so that MD.& PhD reads as credentials. It also admits a leading or trailing & or ; on either side of the period, which looks incidental:
>>> parse("John Smith (MD.& PhD)").suffix
'MD.& PhD'
>>> parse("John Smith (;.MD)").suffix
';.MD'
>>> parse("John Smith (&.MD)").suffix
'&.MD'The question: should a credential with a stray edge mark still read as one, or should only the MD.& PhD shape be admitted? Narrowing it would move readings, and nobody has measured it over the corpora yet. That count comes first (snapshot.py --local for ctgov, which holds hand-typed bracketed credentials). Found by #1153's simplify pass.
Activity
derek73 commentedon Nov 12, 2026
A data point from #1268's work, measured on master 670b349, 2026-11-12, over all 26 corpora (the 16 local corpora as cached).
This issue's shape (a
;or&against a period at a credential's edge,;.MD,&.MD):&.words with no credential chunk, which read as a given name.&after a period (MD.& PhD): ctgov-officials 6, FEC 1.A semicolon inside brackets is rarer still: ctgov 2 names, ORCID 2. Today's readings of made-up names in the same shapes:
Jane Doe (;.MD);.MDJane Doe (&.MD)&.MDJane Doe (MD.;PhD)MD.;PhD(one word, through its chunks)Jane Doe (MD; PhD),(MD;PhD),(MD ; PhD)Jane Doe (M.B; B.CH)M.B; B.CHJane Doe (B.Sc; M.Sc; Ph.D.)Doe B.Sc;, familyM.Sc;, suffixPh.D.Jane Doe (B.Sc;M.Sc;Ph.D.)Doe, familyB.Sc;M.Sc;Ph.D.Jane Doe, MD; PhDMD; PhD(rules/C10)What the rows show:
John Smith (MD,PhD)read the credentials when the comma has no space beside it? It gives nicknameMD,PhDtoday #1268). rules/C10 reads it only after the first comma.(B.Sc; M.Sc; Ph.D.)row is the one that harms. It is how one ORCID record writes its degrees. The clause ends in a full stop, so it takes the period escape and is read bare. Outside the brackets, with no comma before it, C10 does not reach the semicolons, and the degrees become the surname._period_chunksmoves no real name: the edge-mark shape holds none. The question here would only change(;.MD)-like inputs.🤖 Generated with Claude Code