Self-assessment

Decidable —
or only judgeable?

Your own requirements · four questions · four ratings

You enter three to five requirements on machine behavior, in your own words. For each you At the end it says which can be decided today — and which can only be judged.

01 — What is assessed

Four ratings.
The fourth is a result too.

A passed acceptance test is a snapshot of a sample: it holds for the section that was checked and for the moment it was checked. This check takes your requirements one by one and asks whether anything can be decided about them at all.

1Directly decidableThere is a document and a detectable reaction.
2IndirectlyCheckable once the requirement gets a document.
3ConditionallyCheckable, but the statement is not sharp.
4Not decidableNo detectable reaction — technology is missing.
—What that meansThe method makes it visible. It does not remove it.
—And what followsA gap you know is a risk. An unknown one is an assumption.

WhiteBox does not mean you can see the code. WhiteBox means conformity becomes decidable instead of a matter of judgment — and that is exactly the distinction this check makes, requirement by requirement.

02 — The four questions

Four questions
per requirement.

They always refer to the same transmission path: is there a sender, is the receiver readable, is the statement sharp, and can it be established without the builder present.

01
Is there a document stating this requirement — not the code?

Is there a sender? Without one there is nothing to compare the behavior against.

02
Is the expected reaction technically detectable on the line?

Sensor, signal, state. Is the receiver readable? Where no reaction is detectable, nothing can be decided formally either.

03
Is it unambiguous when the requirement is met and when it is not?

Is the statement sharp? “Runs smoothly” is not a checkable requirement, “holds position within x” is.

04
Could this be established without the builder present?

Is it decidable or a matter of judgment? This question separates the two — and it is the most uncomfortable one.

05
And how the rating follows

Four times yes: directly decidable. Question 2 no: not decidable. Question 1 no, rest yes: indirectly. Question 3 or 4 no: conditionally.

06
Why four and not twelve

Because every further question lowers the completion rate without sharpening the rating. Four are enough to separate.

03 — How to read the result

A table of your requirements,
with one sentence of reasoning each.

The result is not an assessment of your line but a sorting of your own sentences. The Behavior Reference (BR) is not after-the-fact documentation of the code. It exists before the code — and the requirements you enter here are the material it would be made from.

What you get

Your requirements with rating and reasoning — as a report to keep.

The not-decidable first

It appears at the top, not the bottom. Decidability presupposes technology: where no reaction is detectable, no model helps.

What the result is not

No statement about whether your requirements are the right ones. Verification asks: was it built right? Validation asks: was the right thing built? The second question stays with you.

Your texts belong to you

The requirements you enter are not analyzed and not used for other purposes. They appear in the report and nowhere else.

04 — Request the report

The report comes
to your address.

The rating appears on the page immediately. The report with your requirements and the reasoning we send by email.

HubSpot form — request the report
What the method does not do

Decidability presupposes technology. Where no reaction is detectable, nothing can be decided formally either — the method makes that visible, it does not remove it.