The Vision

This project was built with a clear ambition: a polished desktop save editor that understands the data it's presenting — not just showing raw bytes. The owner wants people to be able to open a save file and immediately have fun with it, guided by smart defaults.

Randomization Feature (High Priority)

The most exciting feature even in its early buggy state. Goals:

Place player in a random but playable new-game state
Random map, random Pokemon team, random name, random starter pack of items
Constrained to avoid glitches: no glitch items, no glitch Pokemon, no broken maps
Always includes an "escape Pokemon" (HM slave with relevant field moves)
No progress: no beaten trainers, no events completed, clean start
Playability is the constraint — randomization within those bounds

Quote from session 1: "even with this incredibly early feature it was so fun to play I found myself re-rolling new save files even throughout the bugs."

The spirit, in the owner's own early-playtest words (from NewGoals.md): being "dropped into Pokemon Tower as Paddles with a ghost Pikachu named Wolvoom that knows the wildest of moves," navigating out to start the adventure, "only to encounter a level 5 Zapdos along the way and the whole map being darker than normal." Curious, surprising, a little broken-feeling — but the kind of thing that makes you laugh and re-roll. That feeling is the design target.

Map Editor (Medium Priority)

This is one of the reboot's headline ambitions (from the original NewGoals.md). A Gen 1 map carries hundreds of variables — connecting maps, warps, sprites, positions, music, wild Pokémon, tileset, and more — and the goal is to expose all of it without complicating anything: let users do far more with "the area you're in" than was ever possible before.

Imagined UX:

Pick any map in the game and click "load player here" to drop your character onto it.
Edit connections, warps, sprites, positions, music, and wild Pokémon — "editing options you probably didn't even know about."
The app does the grunt work in the background — sensible default warps/sprites/connecting data and all the pointer calculations — so it just works...
...while still exposing full power to anyone who wants to go all the way. Smart defaults, not locked doors. The intent is to encourage the user to go out and have fun with the map.

Full Screen Coverage

Snapshot as of session 13j (see status.md for the live, authoritative list):

Working / data flows + persists:

Font keyboard (filters, search, str propagation, DetailView)
Trainer Card (name + rival render/persist, ID, money, coins, starter, badges, playtime)
Pokémon box screen (sprites/levels, hover name+pen, click → details, no crash)
Pokédex (toggle/click)
Bag / Items (read + edit + count)
Pokémart (looked correct)
File open/save

Working but needs polish (mostly Qt 6 Material-height layout — see status.md open issues):

Pokémon details editor (opens + data persists; re-test moves/DV/EV after the s13h GC fix; level box + box heights still need layout work)
Full name editor (right-side tileset preview blank; description not showing; empty rows when items unchecked)

Not yet exercised/verified end-to-end:

Rival data (name shows; verify starter saves)
Maps / Map Details (verify appBody.push navigation)
Pokémon storage PC boxes beyond party
Hall of Fame

Technical Ambitions

The owner wanted this rewrite specifically because the JS/Electron version couldn't be extended further without becoming unmanageable. Qt gives:

Fast builds (compared to 45-minute Electron builds)
Native performance
Full access to game data through the C++ DB layer
A path to features impossible in the browser sandbox