Choosing an eligible representation

This overview explains how required output, live support, revisions, reuse, and lifecycle narrow the choice of representation. Width alone cannot select a winner for callers with different outputs and reuse obligations. A caller asks for the full vector of a four-input rule, starting from fresh state. There are sixteen assignments, and every output must be delivered in the specified order. Eligibility is determined by that complete requirement. An eligible route can use different internal representations, but it must produce all sixteen exact answers. A method that only provides one point answer has not completed this request, regardless of how quickly it returns. Now change the request to a projected count from a resident structure. The artifact is a number of distinct selected answers, and preparation may already have occurred. This task can have a different eligible set and cost balance from the fresh full-vector request. State the required output. Then ask about live support, repeated queries, revisions, exact operations and complete lifecycle costs. These questions narrow eligible methods before measurement chooses among them. Compare two requests. The first asks for every output of a four-input circuit once; it needs an exact complete vector, and setup is charged once for one result. The second asks for distinct valid feature products under many restrictions while a structure stays resident. It may benefit from a count-oriented representation, but must not count auxiliary witnesses as extra products. Width alone cannot choose either winner. A reduced output, a resident artifact and a reload have different obligations. The current retained evidence keeps learned selectors unpromoted and some native paths guarded. Use those findings as constraints on a proposal, not as a substitute for a new caller's measurements. First decide whether a method can deliver the required exact artifact under the input and resource constraints. This is an eligibility check, not a claim about which eligible method is fastest. A selector may predict which eligible method is cheaper. That prediction has its own feature, dispatch, and possible conversion costs. Keep those inside the request if the caller must pay them. Finally, compare the actual complete task under a declared protocol. Eligibility, prediction accuracy, and measured performance are three separate reports. A correct prediction about one internal kernel need not identify the cheapest complete request. Choose eligible exact paths for the actual requested task. Then compare their complete costs under a matched protocol.