Deletion Semantics

[thelmick-unicon]

Goal: Align on a predictable deletion method that avoids accidental data loss while still allowing for cleanup.

Context: As we design a process for modifying existing data models by uploading a schema, we need clarity on what it means to remove something. Specifically, there is an important distinction between removing an entity or attribute from use and deleting it.

Question: When editing a schema via upload, if an entity or attribute is removed, what should happen by default?

• Should it remove only the association or placement of the entity/attribute, or should it delete the underlying entity/attribute and definition?
• Should an upload be allowed to delete an entity or attribute? If so, under what conditions?
• Should ownership of the entity or attribute affect the deletion behavior?
• When an entity/attribute/association that is part of a transformation is deleted, what happens to the transformation?

[thelmick-unicon]

AI Summary of feedback from 2/13/2026 Wallet Users Community Call:

From Duncan Cox

• Reinforced that deletions should not eliminate attributes entirely.
• Suggested removing only the association or usage within a specific context rather than deleting the element globally.
• Highlighted the importance of future reuse, particularly across evolving use cases (K–12, workforce, informal learning).
• Supported versioning as a safer alternative to deletion.

Underlying theme:
Deletion should be reversible or non-destructive. Elements may not be needed now but could be needed later.