Simplify/standardize obsolete-class modeling

[jclerman]

Issue

Currently, obsolete classes ("terms") in RadLex are modeled in two ways, which are currently not uniformly synchronized:

• Subclass of http://www.radlex.org/RID/RID15849 ('Obsolete RadLex Term')
• Has a value for http://www.radlex.org/RID/Preferred_Name_for_Obsolete (annotation property)

The intent of the annotation property is to make obsolete terms less-likely to be confused with/used as live terms. In practice however, the annotation property adds a maintenance burden and should not be necessary if we stipulate that term-searches should exclude subclasses of 'Obsolete RadLex Term'.

Proposal

Obsolete terms in RadLex will be defined solely by being under 'Obsolete RadLex Term'.

• move all values on preferred-name-for-obsolete to rdfs:label
• any conflicts with existing labels get resolved by adding "(obsolete)" to the label
• synonyms should be moved to existing terms when a suitable one exists
  • however, fully deprecated concepts may keep synonyms
• reason-for-deprecation annotation property should be added; existing comment values with that info should be migrated to that property

[jclerman]

This was originally discussed in the RadLex GitHub meeting on 9/29/2026 and the proposal was adopted at that time; implementation is the next step.

[hoodcm]

I support this. Here are the counts for the work, with PR #9 applied.

• 940 classes are obsolete. 912 have no English label, so their obsolete name becomes the label.
• 70 obsolete names match a live label, so they get "(obsolete)".
• 2 pairs of obsolete classes share a name. Adding "(obsolete)" leaves each pair with the same label.
• 46 obsolete classes have synonyms to review.
• 894 have a comment we could move to the reason field. The most common is "duplicate."