work in progress
Delivery 2 min Diagnostic Scope: an inquiry, alone or shared

Signing off a spec does not make it right

The symptom

A document describes the expected result and the constraints. It has been signed off, but a new observation seems to contradict one of its assumptions.

The signal

Separate what is required, what was verified, and what is still assumed. Have the contradiction examined before you change the agreed work.

What’s going on

A specification helps coordinate work and keep decisions. Reviewing it can bring new knowledge: a forgotten constraint, a use case, an extra check. Sign-off does not, however, guarantee that every assumption will survive building and use.

Not everything is a matter of preference. Some requirements correspond to a commitment, a protection, or an operating constraint. They do not disappear because one attempt produced different feedback. You have to understand their reason and identify who can authorise a change.

A document specifies a choice of time slot. The team assumes it will make signing up easier. A first attempt shows that some people do not understand the times offered. That feedback may call for a better explanation, another choice, or more observation; it is not enough to conclude that nobody wants a choice.

A discovery does not grant a right to diverge quietly. Present the facts, their reach, and the options to whoever owns the remit. For a shared document or an external commitment, have the change and its consequences for the deadline and the cost confirmed. Someone starting out can bring a precise case without having to resolve the whole contradiction alone.

Check this

In a working document, add an assumption that matters to the decision:

We assume that …
We will check it by …
If what we observe contradicts that, we will examine … with …

Choose a proportionate check and say what its limits are. A qualitative observation can be enough to reveal a difficulty; a numeric threshold needs a reason and a context.

When the feedback comes, note what was learned and have the decision updated if needed. An assumption that was confirmed is worth keeping too.

From where you sit

To discuss

Which assumption in a document deserves checking, and which requirement has to be understood first?

Discuss this