When a Body Becomes an Inventory
Human Atlas changes its spatial contract as the reader moves the explosion control. The body begins as an anatomical context, then becomes a set of separated structures, and finally approaches an inventory of visible pieces. This chapter explains how that change is built: which parts are laid out, how the layout avoids overlap, why camera controls switch from orbit to pan, and how the viewer protects a tap for inspection from a drag, pinch, or canceled gesture.
Separation begins with the visible subset
An exploded view is often described as an animation, but Human Atlas treats it as a change in spatial representation. The scene does not lay out every source mesh every time. It first determines which parts should be visible from the current system filters, selected set, and isolation state. Only that visible subset is given a packed layout. Hidden parts keep their body-centered locations and receive no visible state, so the inventory grows or shrinks with the reader's chosen scope.
That choice makes system controls and explosion composable. A reader can show only the skeleton and then separate those pieces, or keep several systems visible and inspect their combined inventory. The layout is not a fixed poster prepared independently of the scene. It is computed from the parts that matter in the current state. If a selected structure must remain visible even when its ordinary system is hidden, selection can add it back to the visible set; if isolation is active, the selected set becomes the complete visible subset.
The movement itself has two stages. At lower separation, parts move away from their body-centered positions into system-spaced positions. This preserves some sense that the pieces are leaving an organized body rather than teleporting directly into a grid. At higher separation, the scene interpolates toward a packed inventory layout. The transition therefore changes gradually from anatomical context to inspection surface, giving the reader time to see that the representation—not the underlying identity—is changing.
The visible-set decision also determines which exceptions the scene must preserve. A selected part can remain visible even when its system is not in the ordinary visible list, because selection is meant to keep the reader's current focus available. Isolation narrows the rule further: the selected parts become the visible set, and the surrounding anatomy is suppressed. The layout therefore follows an order of attention. First it asks what the reader has chosen to see; then it arranges that choice. It does not arrange everything and hope that the camera or opacity controls will make the result readable afterward.
This is particularly important for a multi-piece concept. If a search or direct selection names a structure represented by many parts, the explosion control can reveal those elements without requiring the application to invent a single shared bounding box or a single center point. The semantic selection remains one reader-facing choice, while the inventory can expose the geometric members one by one. That is how the view supports both “show me this structure” and “show me what this selection is made of.”
The distinction matters when a part is selected during the transition. Its ID, bounds, concept relationship, and selected state do not change because its on-screen offset changes. The scene can move the same source part from its anatomical location to a visible cell while keeping it available to the same details and picking paths. What changes is the answer to “where should this piece appear for the current task?”
Packing bounding boxes into readable cells
The layout starts from each visible part's bounds. It uses the width and height of each part's anatomical extent, applies minimum dimensions and a small margin, and treats the result as a cell that must receive that part.