How photoArchive became a product.

Problem: Develop had serious foundations, but photographers also need comparison, proofing, detail controls, saved working views, and a timeline that makes edits understandable after the fact.

Change: Dev7 shipped film UI, reference compare, lens calibration detail, soft proofing, XMP write-back, saved views, and timeline work.

Effect: Develop now reads less like a promising panel and more like a working room: compare, correct, proof, write back, and return to a view without rebuilding context.

Problem: A photo archive that can find and rank images still leaves a gap if every meaningful edit has to happen somewhere else. The editing surface needed history, presets, color, geometry, and non-destructive behavior.

Change: The Develop work brought in non-RAW editing, color wheels, transform and upright correction, presets, history, virtual working concepts, and XMP import/export.

Effect: photoArchive became more than a finder. It can carry the image from selection into craft while preserving the local-first archive model.

Problem: The product had many promising surfaces, but too many paths can make a tool feel like a lab. The working desktop needed a default place where search, review, proofing, and publishing made sense together.

Change: The interface consolidated around one desktop default while three-engine search, stacks, private proofing, and publishing became part of the same product story.

Effect: The archive started to feel less like separate experiments and more like a workflow: find the set, review it, make decisions, and publish selected output without exposing originals.

Problem: A private photo archive is much less useful if it only works at the main workstation. The phone experience needed to feel intentional, not like a cramped desktop page.

Change: The mobile work moved from companion access into an installable PWA direction, with timeline browsing and phone-scale controls for the same self-hosted archive.

Effect: photoArchive became pocketable for browsing and showing work while keeping the archive private and self-hosted.

Problem: A ranking loop can identify favorites, but it does not solve the larger photographer problem: browsing a real archive, staying oriented, and finding images by what they are.

Change: The project became photoArchive: a justified library grid, folder context, metadata filters, loupe review, and early AI-assisted search moved the product beyond ranking alone.

Effect: The system gained a recognizable product shape. It could act like an archive, not just a preference experiment.

Problem: Pairwise choices are fast, but a growing archive needs broader context. The old desktop loop needed to become more visual and more accessible.

Change: The web era kept the simple taste question and expanded it into mosaic ranking, where multiple candidates could be compared in one pass.

Effect: Refine became easier to understand at a glance, and ranking started feeling like a product workflow rather than a utility script.