The Documentation Completeness Test for the CCAR-P Exam

A test defined by what a reader can do

Having built a decision log and a control register, how do you know the documentation is actually complete? The CCAR-P exam gives an analyse-level answer: apply a behavioural test. Can a competent architect who was not in the design sessions make a safe change to the system after reading the document? If yes, it is complete; if no, it is not, no matter how long or detailed it is.

What makes this analyse-level is that it is not a checklist. Completeness is not "has a decision log, has a control register, has a diagram." It is defined by an outcome, whether a specific reader can safely act, and a document can contain every expected section and still fail, because passing depends on what the sections actually enable rather than on their presence.

What passing actually requires

Because the test is about safe action, it demands the specific things a reader needs to act safely. Decisions have to be dated, so the reader knows the context and sequence in which they were made. Assumptions have to be labelled explicitly as assumptions rather than embedded as if they were facts, so the reader does not build on an unconfirmed premise mistaking it for solid ground. Open items have to carry an owner and resolution criteria, so the reader knows what is still unsettled and who to ask. These are not decorative; each one is a thing a reader must have to avoid making an unsafe change. A document missing them can be read but not safely acted on.

Why a diagram alone fails

A detailed architecture diagram is the classic document that passes a checklist and fails the test. It shows what the system is, every component and connection, and it looks authoritative. But it does not show which parts are load-bearing: which choices were made to satisfy a hard constraint and which were mere preference. A reader armed only with the diagram can see the shape of the system and cannot tell which pieces are safe to change, so they cannot make a safe change with confidence. The diagram fails the behavioural test precisely because "what the system is" is not "what is safe to touch." This is why documentation needs the decision log's rationale alongside the diagram, and why the handoff failure so often traces to a surviving diagram with the rationale missing.

Length is not completeness

The other analyse-level trap is equating detail with completeness. A long, dense document feels thorough, and thoroughness feels like completeness, but the test does not reward volume. A hundred pages that never label their assumptions or date their decisions can still leave a reader unable to act safely, while a shorter document that does those things can pass. Judge the document by the behavioural test, not by its weight.

What the exam trips candidates on

The first trap is judging documentation complete because it covers every component in the architecture diagram. Coverage of the diagram is not the test; the test is whether a reader can safely change the system, which requires rationale and load-bearing information the diagram does not carry. The credited answer applies the behavioural standard.

The second trap is assuming a document passes the completeness test simply because it is long or detailed. The exam offers an impressively detailed document that still lacks dated decisions or labelled assumptions, and rewards recognising that detail is not the same as the specific things a safe change requires.

Common misreadings to avoid

How this shows up on the exam

Questions give you a document and ask whether it is complete, often contrasting a detailed diagram with a document carrying rationale. The reliable reading is to apply the behavioural test, could an outside competent architect make a safe change from this alone, and to check for dated decisions, labelled assumptions, and owned open items rather than crediting coverage or length.

This test integrates the decision log and control register into a single standard and embodies the returning-architect reader from the three readers of architecture documentation. Failing it is exactly the situation diagnosed in diagnosing rationale loss in a handoff failure.