Making loot announce itself

In short: Dropped loot now tells you what it is before you walk over: the beam means quality, a gem glints in its own colour, and a soft glow and a trace up a sharpened edge mark the rest. By the last pass I called it perfect.

Five passes between 26 and 27 September went on one question: when something drops on the floor, what does it tell you before you walk over to it? By the end I said "Both are perfect where they're at". Getting there changed the rule for what a beam means, and it turned up glints that flew 48 cm when they were meant to hover.

A beam means quality, and nothing else

Pass 18 started from the grid. Items there now sit on a backdrop in their quality's colour, and I said "I love knowing what each item quality is." The beams on the floor said nothing about rarity yet, so I asked for the beams to show rarity, and for gems to take one general light.

A dropped item's beam is now its quality's colour, taken from the same table as the grid's backdrop, so the floor and the grid can't drift apart. Coins lost their beam entirely. The coin gold (1.00, 0.82, 0.35) sat 0.07 from Storied's (1.00, 0.76, 0.32), so a pile of coins would have looked like the best item in the game. A test now holds every pair of beam colours at least 0.15 apart, and all seven of its cases were watched going red.

The numbers still missed something. Rendered side by side from 9 m and blended the way the game draws them, Plain and the gem light came out only 0.035 apart. On paper the pair was fine, and in the picture the gem light looked like a paler Rare.

Gems glint in their own colour

Pass 19 answered it. I could tell a gem from a Plain item, "at least that it is a gem", but not its colour at a distance. So I asked for the gem's colour on the gem, with beams only ever meaning quality.

Gems now throw glints in their own tint. They borrow the coin's glint particles, through a copy of the beacon material with its base colour set to white, because the beacon is gold and would have turned a blue gem's glints muddy. A gem's beam now shows its grade on the item ladder: Chipped grey, Flawed green, Normal blue, Perfect gold. Pass 20: "working great", and the colour read at 5 m.

A glint that flew 48 cm

Pass 20 also asked for the glints to stay closer in. Measuring them showed why. The emitter's circle shape launched every glint outward at 0.26 m/s, so after two seconds they were 48 cm from the gem. That wasn't sparkle around an object, it was a spray. Now they launch at zero and rise at the same speed, and the circles shrank (coin 0.17 to 0.12 m, gem 0.06 to 0.04 m). Pass 21: "fantastic". Then I asked for "just a bit of the sparkles moving slightly away ... I think that would be magical." They now drift outward at 3 cm/s, at most 6 cm past their circle, which keeps a gem's glints inside its glow.

From a beam to a glow

I also wanted the beam's bottom rounded off, "more of this ethereal highlighting of the object rather than an actual beam of light", and the weapon always in front of it. Up close, the blade had been clipping through the light.

The beam is now a glow centred on the item: brightest at the item's height, fading both ways. An item hovers at 0.75 m, so the light is gone 0.15 m above the floor. It also stands behind the item from wherever you are, pushed back along your view by the item's farthest horizontal reach. There are tests for the glow's profile and for a dropped arming sword and greatsword seen from four sides, and all were watched going red. The first run of the "behind" test measured from the wrong camera, because in edit mode the main camera is the scene's, not the player's. It now sets the eye directly.

Pass 21 was honest about it: "I saw the beams end, but I didn't really see a soft glow." Then: "That will do for now, it looks good." I think I know why. The brightest band is at item height, which from standing is below the eye and behind the item. It may simply not read as a glow. That's a guess, and I've left it alone.

The whetstone's edge

A sharpened blade has its own tell. In pass 19 I asked for the glow to trace up the edge, reach the tip, pulse, and repeat. That meant cutting each edged weapon's edge line out of its mesh into a table, so the glint knows where the edge is. The khopesh gets its one cutting edge, and the rest run up both at once. Pass 20: "looks great. Works perfectly." But I wanted "a reminder that it has been sharpened, not a constant reminder," so it went from every 2-4 s to every 6-10 s. Pass 21 said "a tad faster", so now it runs every 5 s, stretching to 8. Pass 21 was also the first time the log proved the glint runs up the edges, not down the middle.

A flaky test, chased

Before pass 22, the glint test failed once. The farthest glint measured 5.2 cm against a 6 cm bar, which is the check that the drift exists at all. Re-running would probably have passed. Instead Claude found the cause: the particle system picks its own random seed, and one unlucky roll put every glint near the middle of its circle. The test now fixes the seed. Five runs were green, and it still goes red with the drift removed. The game code didn't change. A test that fails once and passes twice was going to fail again the week I stopped believing it.

Still open

The glints are hard square pixels. The beacon material has no soft texture, and coin glints have looked like that since early September. I haven't raised it, which may say it doesn't matter or may say nobody looks. The whetstone keeps a slate-coloured beam that isn't a quality, so it breaks "beams mean quality". That one is my call, and I haven't made it. Everything else here is worn and settled, and shouldn't be retuned without another wearing.