paddyg wrote:That's it: we should simply specify pre-babel formats. All images, 3D models and other support files are to be exchanged in Akkadian cuneiform.
Ha! Yes! Of course, why didn't I think of that? [Mel Brooks as Moses]: "I bring to you from God, The Fifteen [CRASH! as he drops a tablet] ... I bring to you the Ten, TEN Commandments!"
The demos have definitely evolved rather than being the result of Intelligent Design. They randomly contain examples of how to do the same thing in different ways! Here
's a recent attempt at ordering them by complexity.
I'm not complaining mind you, as these are much more useful than what is typically provided for free, and even in many commercial offerings.
paddyg wrote:The main issue I would foresee for your fly-through is the scale difference which leads you straight into transitioning from one scale to another and loading resources in background threads. There might be four scales: 1 light year, 1 astronomic unit, 1 million m, 1 m (1e16 : 1e11: 1e6: 1) and though the Shape.scale() method could cope with fitting an Earth sized sphere onto the 1m scale the difference in height between middle and edges of the equatorial faces would be 60,000m so positioning would be a nightmare. Maybe the tidiest transition would be to alpha blend from one to the next.
Fortunately, the astronomical (pun fully intended) scale of how far apart things are in space means that we don't need to display more than one star, planet, moon, asteroid, comet, etc., in any detail at one time. The shot of Earth from the Cassini Saturn orbiter through the planet's rings a few months ago shows just how tiny and far away our "pale blue dot" (as Carl Sagan noted) really is from the gas giants (at least 1.8 billion miles from Saturn). BTW, I've been playing around with the Points you guys provided to create galaxies so that they can be navigated by spinning and flying through them, and while 100,000+ Points isn't quite the same as the ~100,000,000,000 stars in a typical galaxy such as the Milky Way, it's "close enough for engineering approximations", as the mathematician vs. engineer joke goes. AIUI, we only have accurate enough distance information on a few hundred thousand stars to need to worry about their relative 3-D positioning, so everything else can just be plastered in an image file that's wrapped around the inner surface of a sphere at a radius of infinity minus del (the largest distance that can be represented in the number of floating-point bits available). I've been able to use color shades on Points to make smaller or more distant stars appear with appropriate brightness.
paddyg wrote:For 3D model format I suggest sticking to the wavefront .obj as lots of programs can export these (and we've already got the code to parse them into pi3d!) and image formats .png or .jpeg as these definitely work with python imaging on the pi (pil or pillow)
Yep, and I need to keep collecting freeware objects either in Wavefront, JPEG, PNG, etc., formats or something that can be transformed into those formats freely.
The best things in life aren't things ... but, a Pi comes pretty darned close!
"Education is not the filling of a pail, but the lighting of a fire." -- W.B. Yeats
In theory, theory & practice are the same - in practice, they aren't!!!