THE
LAB

Can a compact history view show recorded trends without turning missing samples into healthy-looking values?

A polished graph should not make absent evidence look like a successful measurement.

Synthetic histories and unavailable responses check chart rendering, input validation and repeated refresh behavior.

Fixture-based chart and state tests render stored trends while preserving unknown and unavailable states. Invalid history is rejected rather than drawn as real measurements.

PROVEN

The interface distinguishes recorded values from data gaps and failure states.

Continue testing sparse histories and longer timelines without publishing operational readings.

NONE

Can a publishing workflow generate inspected image copies while retaining the source photograph?

Publishing an image should be an explicit editorial action with a tested output.

Fixture-based lifecycle and image-inspection tests exercise output generation, failure handling and source preservation.

Isolated publication tests cover metadata-stripped JPEG output, publication edits, removal and repeat publication with the source retained. This verifies the tested workflow, not every historical gallery asset.

PROVEN

The workflow has a repeatable test boundary for publication copies and their public descriptions.

Finish end-to-end release validation before treating the workflow as a fully integrated live gallery feature.

NONE

Can both comparison panels support the same configuration controls without sharing selection state?

An uneven comparison interface can make equivalent choices look different.

Regression tests cover same-weapon parity, independent selections, clear/replace behavior and empty/populated rendering.

Either side can configure or clear a weapon independently. Replacing a selection removes its stale attachments instead of reviving a previous setup.

PROVEN

Both panels use a consistent configuration path while retaining independent state.

Continue comparing UI behavior after changes to the game-tool catalog.

Can a game loadout builder and reference view agree about attachment compatibility?

A tool should distinguish a known game rule from a qualitative description.

Focused deterministic tests cover compatible slots, rejected combinations, removal, per-weapon independence and provenance.

Compatibility and reference selectors share one model. Unsupported combinations are rejected; qualitative effects do not become invented exact numerical bonuses.

PROVEN

Loadout configuration and reference browsing use the same compatibility decisions.

Recheck source data when the game changes; passing application tests does not independently validate every live-game value.

Can a mobile horizon view connect sky position, forecast coverage and observing windows as time changes?

A timeline is easier to interpret when direction and altitude are visible together.

DST day bounds, normalized positions and missing/stale weather were tested. Desktop and phone browser checks cover the slider, seven-body display, mission switching and zero ephemeris requests while scrubbing.

Explore shows Sun, Moon and planetary positions across a noon-to-noon panorama. Time scrubbing updates positions locally; hourly forecast gaps stay unknown and cannot display GO.

PROVEN

The horizon and existing compass support the same selected interval, with textual readings and keyboard/touch controls.

Gather field usability feedback. The flat panorama is not a camera projection; cloud shading is illustrative rather than a spatial cloud forecast.

Can five planets be evaluated locally against an independent ephemeris reference?

A local calculation supports responsive planning without an ephemeris request whenever the time changes.

The pinned Astronomy Engine model was compared with airless topocentric reference fixtures. Maximum observed grid errors were about 0.00252° altitude and 0.00565° azimuth.

Mercury, Venus, Mars, Jupiter and Saturn are evaluated locally. All 35 saved JPL Horizons comparisons meet the preselected 0.05° altitude and azimuth tolerance.

PROVEN

The Planets mission selects by maximum altitude and qualifies low-altitude eligibility. Selection is not a brightness or viewing-quality ranking.

Extend independent reference coverage before broader accuracy claims. Refraction, terrain, phase, magnitude and solar elongation are not modeled here.

Can observing plans compare time and place without letting one overall score obscure the selected mission?

The best place or hour depends on the target, conditions and available evidence.

Mission switching, freshness, incompatible forecasts, unsafe conditions, equal atlas cells and missing skyglow were tested. The 1.5 brightness-ratio comparison is explicitly a planning heuristic.

Explore separates hourly mission assessments from whole-night photography results. For supported dark-sky missions, a materially darker site can break a similar-conditions tie; better weather is explained as a trade-off.

PROVEN

Mission-specific eligibility and explanations accompany comparisons. The skyglow preference does not alter numerical weather scores or apply to Moon-focused missions.

Validate the comparison heuristic with field outcomes. Meteor-specific geometry remains unsupported.

Can nearby mapped places be assessed without mistaking a map listing for verified nighttime access?

Place discovery should help exploration while making provider gaps and access uncertainty clear.

Provider normalization, distance limits, malformed responses, outages, map/list selection and origin assessment were tested; documented live probes established a working discovery path.

Explore uses bounded Photon/OpenStreetMap place discovery and can assess the selected origin even when discovery is unavailable. Candidate assessments reuse SkyOps evidence and safety checks.

PROVEN

The map and list share selection; discovery failures no longer remove the ability to assess the chosen location.

Treat listings as leads. Opening hours, permission, after-dark access and local conditions still need checking.

Can a dedicated seeing forecast add optical-turbulence evidence without inventing it from surface wind?

Cloud cover, humidity and equipment-shaking wind describe different effects from atmospheric image blur.

Credential-free fixtures exercise positive/null samples, UTC coverage, stale runs, malformed responses and failure isolation. Existing scores and safety outputs stay unchanged.

The optional adapter and disclosure passed synthetic and built-route checks. The provider remains disabled under the free-only requirement; live payload validation and forecast calibration are outstanding.

NEEDS FURTHER TESTING

Seeing is treated as an independent evidence source. No conversion from surface winds to arcseconds or unsupported confidence percentage is supplied.

Research sustainable no-cost sources and validate their scientific reference, coverage and forecast skill before activation.

Can a light-pollution atlas add useful site context without being presented as a live measurement or Bortle rating?

Astronomical darkness says where the Sun is, not how much artificial light affects a location.

Numeric tile decoding, cell boundaries, missing data, coordinate validation and built-route output were checked; source cells and requested coordinates remain distinct.

The 2025 Lorenz atlas is sampled at the containing grid cell. SkyOps exposes modeled artificial zenith brightness relative to a reference background, with source and spatial limits.

PROVEN

Site background is now a separate, attributed signal. Missing atlas evidence cannot imply a dark site.

Compare atlas context with intentional on-site measurements; local glare, obstructions and tonight’s actual sky brightness remain unmeasured.

Can missing, stale or incompatible inputs stay visible without becoming a confident recommendation?

A numerical score can look authoritative even when its inputs are incomplete or refer to different conditions.

Deterministic provider, evidence, scoring and safety tests cover missing gusts, stale delivery, unit/time/location mismatches and unsafe observing scenarios.

Weather and derived evidence retain source, units, valid time, location, freshness and failure reasons. Missing values stay unavailable; unsafe conditions continue to block recommendations.

PROVEN

SkyOps distinguishes calculated geometry, modeled weather, proxy indicators and unavailable evidence. A conditions score is not a probability of success.

Calibrate planning heuristics against observations before presenting them as measured forecast skill.

Can core visibility, darkness and Moon interference produce a useful photography window?

A dark night alone does not establish that the Galactic core is in a useful direction or above the horizon.

Pinned Astropy/JPL reference cases check window and peak times within seven minutes, altitude within 0.5° and tested azimuths within 3°.

Core altitude is intersected with astronomical darkness; Moon interference and weather remain separate assessments. Independent reference fixtures cover northern, southern and polar cases.

PROVEN

SkyOps now reports a core window, peak direction and explicit reasons when no useful overlap exists.

Compare field photographs with the planning windows. Terrain, obstructions and the extended Milky Way band remain outside this point-direction model.

Can black-hole mass be translated into horizon size, light-crossing time and scale comparisons without hiding model assumptions?

The scale of a black hole is difficult to feel from mass alone; consistent size, time and comparison outputs make the physics tangible.

Mass-unit conversion, linear radius scaling, light-crossing time, inverse-square conceptual density, representative presets and invalid extremes were checked against independent reference values.

The model reproduces the expected scale: 1 solar mass is about 2.95 km in radius, 10 solar masses about 29.5 km, Sagittarius A* spans a million-kilometer scale and M87* is about 128 AU in radius.

PROVEN

A mass input now produces a traceable educational model with CALCULATED, ESTIMATED and EDUCATIONAL provenance kept visible in the interface.

Keep spin and Kerr geometry outside V1 until they can be introduced as a clearly separate model rather than blended into Schwarzschild results.

Can camera, lens, sensor and sky-position inputs produce useful exposure guidance without presenting an estimate as certainty?

Useful exposure advice needs to show what was calculated, what was estimated and which physical or equipment limit shaped the recommendation.

Full-frame and APS-C examples, NPF tolerance modes, declination changes, tracked and untracked cases, pixel displacement and invalid inputs were covered with deterministic checks.

The planner now compares the 500 Rule with simplified and complete NPF limits, calculates field of view and pixel trailing, adjusts for declination and keeps tracked ranges explicitly estimated.

PROVEN

Exposure guidance now has visible provenance and explains its limiting factor instead of offering a single unexplained shutter value.

Compare recommendations with real captures from representative cameras and lenses while keeping the current calculation baseline intact.

Can ordinary distance inputs become an intuitive explanation of how far into the past we are seeing?

Distance becomes more meaningful when it is translated into elapsed light time and the age of the view arriving here now.

Known physical constants, representative objects, unit-equivalence checks, duration boundaries and very large finite inputs were tested independently from the interface.

Cross-unit checks agree from kilometers through parsecs: the Moon resolves to about 1.28 light-seconds, 1 AU to 8 minutes 19 seconds and Proxima Centauri to 4.24 years.

PROVEN

A raw distance converter became a perspective tool that explains both travel time and what it means to observe a distant object now.

Add new reference scales only when they improve understanding without turning the tool into a catalog.

Can SkyOps keep strong astronomical potential separate from unsafe observing conditions while still surfacing a relevant lunar eclipse?

A planning tool is only useful when its final recommendation respects real conditions and does not erase the event a photographer was trying to plan around.

The scoring, conditions, event and report layers were exercised with severe-thunderstorm, unusable-weather, safe-clear and August 2026 eclipse scenarios.

Severe-weather warnings now force a SKIP recommendation even when astronomical potential is Excellent. Verified 2026 lunar-eclipse contacts remain visible with an explicit unsafe-viewing note.

PROVEN

Sky quality, observing safety and event relevance became distinct signals. Unsafe or unusable conditions now gate the final recommendation instead of merely reducing a score.

Continue expanding verified event coverage and calibrating photographic guidance without weakening the safety gate.

Can the Photo Archive serve sharp responsive images without publishing original files or embedded location metadata?

The archive needs to look sharp on an iPhone without exposing precise capture location or full-resolution source files.

Each supplied photograph generated 960 px and 1536 px derivatives. EXIF and image-resource metadata were removed while the ICC color profile was retained.

Ten originals produced twenty responsive WebP derivatives. That initial derivative batch retained its color profile without EXIF or GPS metadata. This historical batch result is not a claim about every later gallery asset.

PROVEN

Future archive additions now have a repeatable public-image policy: responsive derivatives, preserved composition, no automatic location disclosure.

Reuse the same process for future photographs and add only broad, human-readable locations that are intentionally supplied.

Can Shortcuts retrieve the resulting file size of a converted HEIF without modifying the original photograph?

A real output measurement makes storage savings more useful and more trustworthy than a prediction based only on the source file.

The File Size detail was read from the converted HEIF immediately after conversion inside the existing per-photo loop.

The converted HEIF exposed a File Size value in bytes during runtime. The measurement works; the user-facing unit display still needs another pass.

PROVEN

RawDizzzy can now inspect the generated asset. The test also exposed that raw bytes should be normalized before being shown to the user.

Convert the result cleanly to MB or GB and decide whether the final interface should report each file, the batch total or both.

Can an Apple Shortcut convert selected ProRAW images into high-quality HEIF copies, save them to an album and preserve the originals?

Every later storage, measurement or deletion feature depends on the core conversion path being safe and repeatable.

A small selected batch was processed one item at a time, converted to HEIF and saved as copies before any experimental behavior was added.

End-to-end compile and runtime testing passed, including album-specific saving with the original photographs retained.

PROVEN

The working conversion path became a protected baseline. New behavior is tested around it instead of rewriting it.

Keep the baseline unchanged while experimental measurements and user controls are developed in isolated revisions.