How these questions are written

Where an item comes from, what the verification ledger records, what screened and verified mean here and why they are not the same claim, and what happens when a question turns out to be wrong.

A machine has checked this bank. A person has not. None of the 899 questions published here carries a human signature. 898 questions have been read against the clause they cite by an automated pass, which found no contradiction.

Every question page states which of those is true of it, before the answer, in visible copy. That arrangement is what this page exists to describe rather than to soften.

Where a question comes from

Every item on this site is written for it, against the awarding body's own published objective list. Where the objective names a command, it is written against that command's manual page as well. Nothing here is collected from a dump, reconstructed from somebody's recollection of a real paper, or rephrased out of another bank.

Each question carries one citation: the coordinate of the clause it was written against — an LPI objective number, an ITIL syllabus clause, a Scrum Guide section, a manual page. The citation says where to read the clause; it does not reprint it. Syllabus text belongs to the body that published it, and the one document quoted here at any length is the Scrum Guide, which its authors published under the same licence as this bank. Legal and privacy has the full argument and the credits.

Every option carries its own reasoning, including the wrong ones. That is not decoration: for these exams the difficulty lives in the distractors, and a question that tells you only which answer was right teaches you nothing about why the one you picked was not.

What the ledger records

Verification is kept out of band, in a separate file per paper, rather than as a flag on the question. A question cannot mark itself as checked, and a check is a statement somebody makes about a specific wording at a specific time. One record holds this:

contentHash
A hash over the parts of the question a check is about — the stem, the options, the answer key, the objectives. Not the lesson it is filed under, and nothing cosmetic
verifiedBy
Who, or what, did the checking
verifiedAt
When
verifiedAgainst
The source it was read against, by its coordinate in the published document
method
human or machine. The field this whole page turns on
note
What the check actually established, in the checker's own words

The hash is what makes the record worth anything. Edit the stem, change an option or move the answer key, and the hash stops matching: the record no longer applies, the question drops back to unchecked by itself, and its page says so. Nobody has to remember to withdraw a signature after an edit, because the signature was never attached to the question — it was attached to that wording of it.

Screened and verified are not the same claim

Verified means a person opened the question, read it against the source it cites, and signed for it. It is the only thing this site calls verified and the only thing that counts towards any verified figure anywhere on it.

Screened means an automated pass read the same evidence — the stem, the options, the answer key and the cited clause — and found no contradiction. It is deliberately the weaker claim, for a specific reason rather than out of modesty: the same kind of process wrote the question, so it can confirm its own mistake. A screened question is a good draft. It is not settled fact, and this site does not present it as one.

A screened question is published — its page may be indexed and it may appear in a mock paper here — and it carries a notice saying that a machine checked it and a person did not, in the visible copy and in the page's structured data, so that a reader arriving from a search result and a model reading the markup are told the same thing. A question with neither check is kept out of search results and out of mock exams until it has one.

What happens when a question is found wrong

It is retired, and its page stays. The page says the question was found to be wrong or misleading, it leaves every listing and the index, and the machine-readable markup asserting its answer is dropped entirely — a page telling a person the answer was withdrawn while telling a model the answer was correct would be worse than deleting it.

The page stays because a deleted url does not disappear. It becomes a soft 404 that looks exactly like a bug, and answers already given to that question still refer to it. A withdrawn question is more use to a reader than a missing page. 1 question is retired today.

The same thing happens one size up. A whole paper can be withdrawn — taken off the shelf because there is no longer enough of it to sit the exam — and when that happens every url it had keeps answering, with the notice and nothing else: no question, no options, no explanation.

What leaves your browser

There is no account and no sync. Your answers, your schedule and your session ids live in this browser's IndexedDB and nothing in the application sends any of them anywhere. Clearing your browser storage deletes them, and there is no copy anywhere else.

Page views are counted, by analytics software running on the operator's own server, with no cookies and no third party involved. Legal and privacy states exactly what is recorded, on what legal basis, and how to stop it.

The rest of it

Every exam published on certotter, and the questions under each of them.