I love water zoom. It is pretty and direct.
It may, however, give a mistaken impression of the size of atoms. Each zoom is about 5x, so the total magnification is about 3000x. In fact, starting with a view that might be a km , you would need 17 or 18 zooms of 5x each to get to a scale of about 1 nm, which is the approximate width of the final picture. If you increased the magnification of each zoom to 100X, then you could reach atoms in six steps. That would mean magnifying just one pixel to fill the screen each time.

I (Bob) imagine that our learning activities would be constructed from only a few different kinds of components, because each kind of component needs to be very special. A minimum list of components might be: a grapher, a text display box, a graphic box, the Molecular Workbench, a talking character (what I incorrectly called an avatar), a question tool, a drawing tool, a report writer, buttons, and sliders. That's a lot, but I cannot imagine less. The reason to keep this list short is because each needs to be able to be "UDL-ized", by which I mean that its appearance and rate of change needs to be alterable and it has to describe itself in text or speech. We probably also need a background that can be changed.