Structure types: inspection questionnaires and numbering

A structure is a fixed asset on your network: a bridge, a boardwalk, a culvert, a sign, a jump. Unlike a task, it is never finished — it exists, it ages, and it has to be inspected.

The structure type is what makes that inspection possible. It carries three things nothing else in Sentinel carries:

• a questionnaire — the questions you ask standing in front of the structure, with their relative importance;
• automatic rescheduling — how long until you come back, based on the score;
• numbering — the name pre-filled at creation: PV-001, PV-002, PV-003.

It's the richest section in the configuration, and by far the one that pays off most once it's set.

Where: left navigation bar, under Administration → Structure types.

Two families, same as tasks

Four types are installed when your place is created: Bridge, Drain, Sign and Jump. You can't rename or delete them; you only set which activities they're offered in — plus their questionnaire, their scheduling and their numbering, which are genuinely yours.

Custom types are yours from end to end.

Activities behave exactly as they do for task types: a type is only offered in the activities assigned to it, a type with no activity becomes invisible, and that's the only way to retire one since there is no delete.

A type's page has four tabs. The first, Basic information, holds only the name, the icon and the activities. The other three are the real subject of this page.

The evaluation questionnaire

This is the heart of it. A questionnaire turns "I looked at the bridge" into a score out of 100 and a risk level, comparable from one inspection to the next and from one structure to another.

Nothing exists until the box is checked. A type without a questionnaire is still perfectly usable: the structure is placed on the map, you can see it, you can attach tasks to it. It simply has no inspection report.

"Copy from" replaces everything. The app confirms before doing it. It's meant for building a second bridge type in thirty seconds, not for merging two questionnaires.

Writing a question

The title is what you see in the list, in the field, small. Short and concrete: "Decking condition", not "Verification of deck structural integrity".

The description is the real instruction sheet. This is where you win comparability between two inspectors: write what the ratings mean. "Rate 1 if a plank gives way, 4 if the decking is sound" beats "Check the planks" a hundred times over. Without that sentence, two people rate the same boardwalk 2 and 4.

Importance has four steps, and they're real weights in the calculation:

A "Very high" question therefore weighs six times a "Low" one. It isn't a decorative ranking.

How the score is computed

In the field, every question is rated 1 to 4, or marked not applicable.

The calculation is exactly this:

score = Σ (rating − 1) × weight   ÷   Σ 3 × weight   × 100

Three consequences that surprise people who don't know them:

All ones give 0, not 25. The scale starts at 1; it's rating − 1 that counts. The worst possible result really is zero.

"Not applicable" removes the question from both sides of the division. It neither penalizes nor rewards: it disappears. A missing handrail on a boardwalk at ground level doesn't drag the structure's score down.

An unanswered question counts as 1. It's treated as the worst rating, not as "not applicable". If a question never applies to a structure, tick "not applicable" rather than skipping it.

The score then falls into a risk level:

The safety question

A question marked Safety question has veto power: if it gets 1 out of 4, the structure is classified high risk, whatever the score.

That's the mechanism that stops the arithmetic from lying. A bridge whose anchors are failing but whose everything else is spotless would post a perfectly decent score — and be declared sound. Mark "Anchors and footings" as a safety question and it goes high risk, as it should.

Reserve it for what hurts someone. Two or three per questionnaire, not eight.

Automatic rescheduling

The tab only appears if the questionnaire is active — it would have nothing to schedule otherwise.

When it's on, the next report's date is computed when the inspection closes, from the risk level obtained. A sound structure reschedules far out; a high-risk one comes back in two weeks.

This is the setting that turns inspections into a cycle rather than a chore you forget. Set it once, and the list of structures due builds itself.

The delay is expressed in days (1 to 30) or months (1, 2, 3, 4, 5, 6, 9, 12, 18, 24). For high risk, use days: a dangerous structure doesn't reschedule in months.

Numbering

In the field, the person placing a trail's thirtieth marker doesn't want to decide what it's called. Numbering hands them the name up front.

The setting is per activity and per type. "Boardwalk in mountain biking" and "Boardwalk in hiking" are two independent series, with different prefixes if you want.

Three fields:

• Prefix — up to 16 characters. PV, BRDG, DR.
• Separator — hyphen, underscore, dot, space, or none.
• Digits — 1 to 10. With 3 you get 001. Choose based on how many structures you expect, not how many you have today: going from PV-99 to PV-100 breaks alphabetical sorting, PV-099 then PV-100 doesn't.

The scope: per trail, with off-trail on its own

This is the most important point in the section, and the least obvious.

The series restarts at 1 for each trail. The first boardwalk on "La Coulée" is PV-001, and the first boardwalk on "Le Belvédère" is also PV-001. Off-trail structures form a separate series of their own.

That's deliberate, and it matches how people talk in the field: "the second boardwalk on La Coulée", not "boardwalk number 47 on the network". If you wanted a single network-wide numbering, this won't give it to you.

It's a derivation, not a counter

Sentinel stores no counter. On every creation it looks at the names of existing structures of the same type, on the same trail, reads the highest number in the series, and offers the next one.

Four practical consequences:

A deleted number gets reused. Delete PV-003 when it was the last one, and the next boardwalk will be called PV-003 again. There's no permanent gap.

The offered name is only a suggestion. You can replace it. The app simply warns you if the name is already taken on the same trail.

Changing the separator restarts the series. Recognition requires the separator to match. After switching from PV-001 to PV_001, the old names are no longer seen as part of the series, and the next name offered will be PV_001. The prefix, on the other hand, is compared case-insensitively: pv-004 does continue on from PV-003.

With no rule at all, Sentinel still continues an existing series. If your structures are already named Bridge 12, Bridge 13, the next will be offered as Bridge 14 without you configuring anything. The rule exists to impose a shape, not to switch the mechanism on.

Finally: unticking an activity turns the suggestion off, it doesn't reset it. The name field will simply be empty at creation.

On mobile

The mobile app carries the type list, creation, editing, and the full questionnaire editor — questions, importance, safety questions, scheduling.

The path: Account → your place → Structure types.

One thing is missing: the Numbering tab doesn't exist on mobile. Mobile uses the rules — the name does arrive pre-filled when you create a structure, including offline, by reading the structures already cached and the ones sitting in the sync queue. But prefixes are only set from the web.

Who can change all this

Structure types fall under the Structure types right in the place settings.

• With read: see the types, the questionnaires and the rules, change nothing.
• With read and write: everything described on this page.

It's a right distinct from structures themselves: someone can inspect structures all season without ever being able to touch the questionnaire — and that's usually exactly what you want, because changing weights mid-season makes inspections incomparable. See roles.

Four pieces of advice

Write the rating scale into the description. It's the single gesture on this page that decides whether two inspectors give the same structure the same score.

Two or three safety questions per type. Beyond that, everything lands on high risk and the classification stops meaning anything.

Freeze the weights before the season starts. A score is only comparable over time if the questionnaire hasn't moved.

Set the numbering before the first outing, not after. Renaming thirty boardwalks by hand is a lost afternoon; ticking a box takes ten seconds.