Part I · Chapter 1 · The Ask · lesson 9
This is a place to think, not a place to keep the only copy
8 min
Part I · Chapter 1 · The Ask · lesson 9
8 min
By the end of this subject you will have typed a boundary, a set of risks, a list of controls with owners against them, and a description of every artefact you could produce. Written down in one place, that is an inventory of your organisation's known weaknesses, sorted for convenience.
It is worth knowing what happens to it, and this lesson is early rather than buried because the honest moment to say so is before you have filled it in.
Predict
One key in your browser's local storage, holding one record per thing you typed: each risk, each control decision, each piece of evidence, plus three singletons for your boundary sentence, your six organisation facts and your period dates. Every record carries its own timestamp, set by whichever browser last touched it.
If you are signed in, those records also go to your account, one row at a time rather than as a blob. That detail exists for a specific reason: a whole-state push would mean the last device you typed on silently overwriting rows it had never seen, and losing an afternoon of somebody's register is not a tolerable failure. Two devices editing different rows both survive. Two devices editing the same row resolve to the later edit, and the losing version is reported rather than quietly dropped.
Deleting something leaves a tombstone — a row with a later timestamp and nothing in it. Without that, deleting a risk on your laptop and then opening your phone would bring the risk back, because the phone still holds a row the laptop can no longer contradict.
There is no score field, and nowhere to put one. The risk editor stores likelihood and impact as bands and a position; nothing anywhere multiplies, sums or averages them, and that is a property of the component rather than a guideline a lesson could talk you out of. Track 2 is the argument for why.
There is no aggregate figure of any kind. No percentage complete, nothing anybody could call a maturity level, and no readiness word with a number beside it. The only count in the whole subject is the number of open contradictions, and it rises as you fill the workspace in honestly — declare a period and the evidence outside it becomes visible; mark a control as operating and the missing artefact becomes visible. It falls only when you do actual work. A number somebody has deliberately made larger cannot be screenshotted and sent to a customer as a claim, which is the whole reason it is the only number here.
Nothing computed is stored. The contradictions are a pure function of your rows, recalculated every time the panel renders. A stored finding would outlive the rule that produced it and become a claim nobody checked.
And no artefacts. The evidence section holds descriptions of evidence — what it is, what it shows, how often it is produced, which period it speaks to. It does not hold the screenshot, the ticket export or the signed review. Nothing you upload goes anywhere, because there is nowhere to upload it to.
This is a place to think. It is not a system of record, and the gap between those two is not a feature list.
A system of record has retention you can state, access control you can demonstrate, an audit trail of who changed what, and a way to give a reviewer read access without handing over your login. None of that exists here. The register also has no ability to be the thing somebody tests: a reviewer looking at your controls needs the artefacts, and this holds only your sentences about them.
So keep the real one wherever your organisation keeps things it would be embarrassed to lose — and treat what you build here the way you would treat notes from a workshop. Useful, yours, and not the copy anybody else depends on.
Check