The current everything-is-a-box rendering sucks. We definitely need to support the rendering (in color) of ACT-R's basic types (rect, circle, line, ??), but should probably implement it extensibly. We'd select the renderer based on one of the visual.type strings defined for the percept. Something like this could actually be flexible enough to render actual snapshots of the object (if visual.image had the RGB data)
The current everything-is-a-box rendering sucks. We definitely need to support the rendering (in color) of ACT-R's basic types (rect, circle, line, ??), but should probably implement it extensibly. We'd select the renderer based on one of the visual.type strings defined for the percept. Something like this could actually be flexible enough to render actual snapshots of the object (if visual.image had the RGB data)