Abstraction and Reuse Decision Examples

Positive example — one policy owner

• Context: several entry points calculate the same governed eligibility rule and have begun to diverge.
• Mode: balanced.
• Decision: give the rule one authoritative owner and keep entry-point formatting local.
• Why it fits: shared knowledge changes together without coupling unrelated representations.
• Rejected alternative: leave policy copies in place because their surrounding code differs.

Negative example — similarity trap

• Context: two workflows currently contain similar field-mapping steps but serve different consumers.
• Mode: conservative.
• Bad decision: introduce a configurable universal mapper after seeing the second shape.
• Why it fails: it couples independent concepts and predicts variation without evidence.
• Better decision: keep both mappings local until common knowledge and variation become stable.