None of this is technically demanding. You could represent a billion stars in the Milky Way with 1 Gigabyte of memory, and they would all twinkle. The galactic arms would stretch out from the galactic Center, and you could navigate, from memory (or with automated aides) to The Sun. A planet could be represented, here, in moderate detail, by a few Kilobytes of data, and so you could approach the Earth, rotating in its orbit, spinning on its axis, accompanied by its moon ... and now there is an abundance of information.
The problem is only in one sense too much information, or too much data. The real problem is how you rationalize it. An Earth could come with ... but you have to be careful not to ask for too much ... pools in rivers, each represented as a plane, or a block (of water), and also chutes down which streams course, represented as tilted blocks, so that the river course can be represented as water, and, in fact, a specific volume of water. Well, pools are embedded, normally, in tilted planes, river bottoms, and river bottoms are embedded in another set of tilted planes, river banks, and river banks are embedded in another set of tilted planes, flood planes, and mountains are embedded in flood planes (but there are streams and stream beds and stream banks embedded in those - in their ridges).
Maybe that's an example of too much. The Surface of An Earth can be represented using topographic layers, and a full set of topographic layers in reasonable detail cannot require all that much memory. I don't know where to get a set of topographic layers - for a brief time I did, but I lost track of the information - but I know where to get "raster elevation data" for the entire Earth. This divides the surface of the Earth into a set of Elevation Pixels. Combine that with River Pool Data, Road Surface Data, and the like, and you could create a quite explorable model of The Earth.
These kinds of geometries are so efficient, when it comes to using Data to represent Geometries Visually that a vast amount of detail could be stored on any hard drive or transmitted over the web in no time at all. Then, too, when you start adding extreme levels of detail it'll probably be for a reason, so the database won't necessarily grow out of all proportion. My computer can handle one of these models, but the readily available software is limited, and less limited software is harder to get - and to manage. It's not that the task is hard to manage, rather it's like a conspiracy to offer people as little of this as possible. Well, I can't write the software myself - I've tried and failed - but I'm exploring what the available options will do. Actually, there's one available option, SketchUp. Well, when your resources are very limited, which mine are, maybe in part because I'm far from being a SketchUp expert, if you want to do anything at all you need to greatly simplify, but is sometimes possible to still produce a meaningful result. I've been struggling with the idea that to represent a whole stretch of country (all of central Arizona) you need to represent a large number of features, and I've concluded that, for my purpose, I actually don't ... need to do that. Instead, I am going to try representing two quite distant places, positioning them in model space as representatively as possible, one, maybe, a city block in Phoenix, and the other a mesa top a hundred miles to the north. Starting places.
this could be called yet another approach ...
