Where is the boundary between pragmatism and negligence in architectural decisions?

The boundary

Reversibility and disclosure.

A pragmatic shortcut is one that is cheap to reverse and recorded — named, dated, owned, with a trigger for revisiting. A negligent one is irreversible, undocumented, or both.

Legitimately pragmatic

• Unoptimised implementation behind a clean interface.
• A manual operational step with a documented scaling limit.
• Simple infrastructure at low volume.
• Deferred features, deliberately listed.
• A compensating control accepted explicitly, with a stated residual risk and a dated commitment to the full one.

Each is a real trade with a known cost and a route back.

Negligent regardless of pressure

• Data model shortcuts, because migration is expensive and the shape propagates to every consumer.
• Skipping idempotency on operations with irreversible effects, which produces duplicate charges and sends discovered by customers.
• Omitting auditability, a property of every write path that cannot be retrofitted without touching all of them.
• Security shortcuts on data handling, where the consequence is exposure rather than debt.
• Undocumented shortcuts of any kind, because they become the architecture by default when nobody remembers they were temporary.

The distinguishing question

"If this turns out to be wrong, what does it cost to change, and does anyone know it was a choice?"

A shortcut with a cheap reversal and a written record is engineering judgement. One with an expensive reversal and no record is a decision made by omission — and the people who inherit it cannot tell it was ever a decision.

The professional obligation

Pragmatism includes saying what was traded. An architect who ships fast and does not record what was skipped has transferred a cost to a future team without telling them, which is the part that makes it negligence rather than judgement.

Speed is a legitimate driver. Treating every shortcut as debt to resist makes the architect an obstacle and is not the alternative being argued for.