Main idea

A backlog item is useful up to delivery. After that it loses most of its value: tickets drift out of sync, so do specs, and the behavior put into production differs from what was written at the start. What lasts is elsewhere — the code, the product's observable behavior, and the documentation maintained from that reality.

So you should spend the minimum useful time in it: enough to collaborate, enough to act, not enough to produce documentation that will rot. The writing effort is calibrated on the artifact's lifespan, and a coordination artifact's lifespan is short by nature.

Why it matters

This settles a false dilemma between writing little and writing well. The question isn't the quality of the writing, it's the horizon it serves: the same care applied to a perishable artifact produces a misleading archive rather than memory.

It also gives the inverse rule for what has to last. A long-lived artifact deserves the effort you refuse the ticket, provided it is rebuilt from what is actually delivered.

Nuances and limits

Some contexts require traceability of intent — regulated industries, audits, public procurement. Writing there outlives its use because it stands as proof, which is a different function from coordination.

And "disposable" doesn't mean useless: coordinating several disciplines around one topic is real work, simply bounded in time.

Open questions

• How do you recover, at delivery, the decisions taken along the way before the artifact loses its value?