Jev Concepts

The three question types

Every Jev question is one of three types. They differ in the shape of the answer, not in how they are sent: you can mix all three in a single request, and the TypeSafe docs note that every question in a request sees the same state and is evaluated independently.

Which type should you use?

Pick the type that matches the shape of the answer you need, not the one that sounds closest to the question. The TypeSafe docs put it in terms of what your code does with the answer: a Choice between `refund`, `rebook` and `information` maps straight onto three code paths, a Score maps onto a threshold, and a Noul maps onto an `if`.

flowchart TD
    A[What shape is the answer?] --> B{Yes or no?}
    B -- Yes --> C[Noul]
    B -- No --> D{Do the answers<br/>have an order?}
    D -- No --> E[Choice]
    D -- Yes --> F{Can you describe<br/>each level?}
    F -- Yes --> G[Score]
    F -- No --> E

One case is worth calling out, because the TypeSafe docs single it out too. "Is this candidate strong in Python?" looks like a yes/no question, but "strong" is a degree. A Noul of 0.5 there means the model gives yes and no equal probability, not that the candidate is halfway up some scale. If you want the degree, write a Score with levels you define; if you want a yes/no, define the condition so there is no middle ground.

Other concepts

• State — the content each question is asked about. It can be plain text or a JSON object, and questions can point at one part of it by path.
• Confidence — the 0-to-1 value Choice and Score answers carry, computed from how the probability is spread. Noul does not have one.
• Probabilities — the full distribution behind every Choice and Score answer, which sums to 1.