{"schema_version":1,"experiment_id":"eoa_inverse_innovation_exp05_complete_proposal_portfolio20_20260803","cell_id":"negative_space_design__astronomy_astrophysics","arm":"COMPLETE_PROPOSAL_PORTFOLIO","candidate_id":"nsd-astro-protected-response-window","proposal_index":1,"version":0,"title":"Protected Response Windows for Time-Critical Transient Follow-up","problem":"A telescope queue can assign every usable interval before a fast-changing transient event is sufficiently localized or classified. The schedule appears efficient, but it leaves no deliberately uncommitted interval near the moment when new localization, classification, or visibility information becomes actionable. Schedulers must then displace approved observations, accept disruptive replanning, or let a potentially discriminating follow-up opportunity pass.","actors":["Telescope scheduling lead","Observatory operations staff","Transient-event follow-up team","Investigators whose queued observations could be displaced","Time-allocation or program-oversight body"],"observable_state":"For each relevant observing period, logs can show the fraction of usable time already committed around forecast decision points, when material event updates arrived, whether an eligible follow-up target was visible after each update, how much schedule churn an insertion required, which observations were displaced, and whether the requested follow-up was obtained before its visibility or scientific deadline.","consequence":"When actionable information arrives into a fully occupied schedule, the observatory may lose the timely observation, incur hurried substitutions, or transfer the cost unpredictably to other approved programs.","affected_objective":"Obtain observations that can discriminate among transient-event interpretations within the target's usable visibility window while preserving accountable treatment of already approved science.","intervention":"Create a bounded, explicitly labeled response window around a forecast information-release or decision point. The interval remains genuinely unassigned rather than being silently treated as spare capacity. Its duration and placement are selected from the surrounding fixed observations inward, and a rule protects it from routine infill. If qualifying localization, classification, weather, and visibility conditions are met, the scheduling lead assigns the interval to the authorized follow-up. If the trigger is absent by a stated release time, the interval is returned to a recorded, recoverable list of eligible observations. The schedule display states why the interval is empty, who may activate it, and when it expires.","structural_mapping":[{"archetype_element":"Attention Competition Map","domain_realization":"Map observations competing for the same interval, distinguishing deadline-sensitive transient decisions from observations that can move without losing their purpose."},{"archetype_element":"Omission Candidate","domain_realization":"Temporarily omit a movable queued observation from the selected interval while retaining it in an eligible, recoverable side queue."},{"archetype_element":"Protected Empty Space","domain_realization":"Reserve an unassigned telescope interval that routine utilization pressure cannot consume before its release trigger."},{"archetype_element":"Positive Form Relationship","domain_realization":"The unassigned interval makes the surrounding fixed program and the later transient assignment jointly legible: the void exists specifically to receive information-dependent observing work."},{"archetype_element":"Absence Boundary","domain_realization":"Specify the interval's start, end, eligible event class, activating authority, and release deadline."},{"archetype_element":"Meaning-of-Absence Check","domain_realization":"Label the interval as intentionally reserved, distinguishing it from missing schedule data, an instrument fault, weather loss, or administrative neglect."},{"archetype_element":"Reintroduction Trigger","domain_realization":"Fill the interval with an eligible follow-up when predefined evidence and visibility conditions hold, or restore a queued observation when the reservation expires."},{"archetype_element":"Accessibility and Recoverability Guardrail","domain_realization":"Keep displaced candidates, reasons, dependencies, and restoration order visible to operators and affected investigators; never hide engineering, calibration, or safety-critical activities."},{"archetype_element":"Clarity or Effect Test","domain_realization":"Use frozen historical event streams to test whether operators interpret and activate the reserved interval correctly and whether it changes deadline-relevant observing opportunities without unacceptable displacement."}],"mechanism_mapping":[{"mechanism_slug":"architectural_void","role":"Treats time as a figure-ground plan: fixed observations form the surrounding solids, while a sized and bounded scheduling void is designed from them inward and protected against later infill.","counterfactual_removal":"Without this mechanism, the interval becomes ordinary slack that successive scheduling decisions can consume, eliminating the intervention's active absence."},{"mechanism_slug":"empty_state_design","role":"Makes the blank interval diagnostically explicit by showing its purpose, status, activation path, authority, and expiry rather than presenting an unexplained hole.","counterfactual_removal":"Without this mechanism, operators may interpret the interval as incomplete scheduling, downtime, or an error and fill or ignore it."},{"mechanism_slug":"editorial_cut","role":"Moves only observations judged recoverable out of the protected interval, preserves their scientific and operational context, and records how they can return.","counterfactual_removal":"Without this mechanism, creating the void could become arbitrary deletion or conceal the scientific cost imposed on other programs."}],"causal_chain":["A forecast decision point identifies when new transient information could materially change the preferred observation.","Movable observations competing with that point are identified, while safety-critical, calibration-dependent, and immovable work is exempted.","A bounded interval is deliberately left unassigned and protected from routine infill.","The visible label makes the absence interpretable and preserves the activation and restoration rules.","Actionable event information can enter the protected interval without first displacing a live commitment.","If no qualifying event appears, the expiry rule returns the interval to an eligible recorded observation.","Comparison with a fully assigned schedule reveals whether protected absence improves deadline-relevant opportunity handling and what scientific cost it transfers."],"baseline":"The observing period is populated in advance with the highest-ranked feasible observations, while late transient information is handled through ad hoc interruption, manual displacement, or whatever slack happens to remain.","nearest_rivals":["A target-of-opportunity policy that permits interruption of already assigned observations","Continuous dynamic rescheduling that recomputes the entire queue after each event update","A preemptible filler program occupying the interval until a transient trigger arrives"],"remaining_contrastive_claim":"Unlike interruption or dynamic optimization, the proposal makes bounded non-assignment itself a protected scheduling resource. Unlike a filler program, it does not create an incumbent observation whose start, dependencies, or sunk setup cost can resist activation; its defining test is whether genuine vacancy at a forecast decision point changes operator behavior or achievable follow-up opportunities.","authority_safety":{"decision_authority":"The observatory scheduling lead may define and activate a trial response window only under rules approved by the existing time-allocation and operations authorities; instrument safety and weather decisions remain with operations staff.","authorized_first_step":"Run a read-only retrospective shadow-scheduling exercise using frozen logs from a bounded set of completed transient observing periods; do not alter any live queue or observing allocation.","excluded_actions":["Changing a live telescope schedule","Allocating new observing time","Interrupting engineering, calibration, or safety-critical work","Withholding the existence or cost of a reserved interval from affected program oversight","Using the reservation for an event class or investigator not covered by the trial rules","Deleting displaced observations or their scheduling context"],"halt_rollback":"Stop the exercise if required timestamps, visibility inputs, or displacement records cannot be reconstructed consistently, or if reviewers cannot distinguish the reserved state from an outage. Because the first step is a shadow replay, rollback consists of discarding the simulated schedules and retaining the unchanged historical record."},"negative_tests":{"strongest_counterevidence":"Frozen-log replay shows that the existing dynamic queue or target-of-opportunity process already accommodates actionable updates before their deadlines with no greater displacement or churn, while protected intervals repeatedly expire unused or exclude higher-value approved observations.","problem_falsifier":"The inferred problem is falsified if the sampled event logs show that actionable updates generally had sufficient reprogrammable capacity available and that failures, when present, arose from weather, instrument capability, source visibility, or decision latency rather than schedule crowding.","intervention_falsifier":"The intervention is falsified for the tested setting if plausible placements and durations of the protected window do not increase the number of deadline-feasible follow-up opportunities, do not reduce disruptive replanning, or impose scientific displacement that the authorized evaluators judge unacceptable relative to the baseline and rivals.","risks":["Reserved time may expire unused while feasible approved observations were available.","Forecast decision points may be too uncertain for a bounded void to align with actionable information.","Programs may compete to classify their events as eligible, shifting rather than resolving allocation conflict.","An explicit blank interval may still be mistaken for downtime or incomplete scheduling.","Recoverable observations may have hidden setup, cadence, or calibration dependencies that make later restoration infeasible.","Protection rules may reduce operator discretion during weather or instrument changes."]},"next_evidence_step":"Select no more than 12 completed transient observing periods with reconstructable schedules, event-update timestamps, visibility constraints, and displacement records. Using only information available at each historical decision time, have schedulers produce a baseline replay, a protected-window replay across a prespecified small parameter grid, and a preemptible-filler rival. Record trigger interpretation, deadline-feasible follow-up opportunities, displaced program time, restoration success, schedule churn, and operator misreadings. Review the comparison before authorizing any prospective or live trial.","prior_art_status":"UNSEARCHED","diversity_from_prior_proposals":"No comparison with prior experiment proposals was performed; this sealed candidate was derived only from the supplied archetype, mechanisms, and astronomy-and-astrophysics domain card.","revision_record":{"parent_version":null,"progress_targets_addressed":["Initial complete proposal","Concrete observable domain state","Causal preservation of protected absence","Bounded authority and first evidence","Explicit rivals and falsifiers"],"conceptual_changes":[],"operational_changes":[],"evidence_changes":[],"claim_changes":[]}}