Finally came up with a labeling system that doesn’t suck. It’s based on trimeshes, and should scale a lot easier. Rather than having to scale and reposition a bunch of separate 3d objects, each one will be a single trimesh. Plenty of work to go, but nothing I don’t understand, which is a huge step up. The entire idea I owe to the amazingly talented Jeff Quan.
UV coordinates are the devil, but I’m getting closer to even understanding those.
I have a good resizing method, and perhaps half of a spacing method. Zooming back in makes things… wonky.
Still, good progress on what is a surprisingly complex little fiddly bastard.
I’m pretty sure it’s going to kill me.
The moo-clone now has a semi-functional zoom window. Everything seems to work ok so far. Planet positioning and whatnot seems to be fairly sorted. So, yay.
With that out of the way I need to have the star labels scale, so that zooming out doesn’t make everything unreadable.
Got the rough sketch in place of displaying stars with their planets. The deal is that when you click a star, you get a new window that shows that star’s detail view. So you can look at planets, orbitals, so on.
It’s all a bit… hackish… but it doesn’t strike me as terrible, so I’ll run with it. All that actually happens, right now, is that you get a new window that then pulls a reference to the last star clicked. Next step is to try and really display it, which may be where this all falls apart. I might end up with everything potentially visible in the main view, with these subviews just there as a convenience. All depends on how wacky the camera gets.
Finally got my star labeling system working properly. I remembered the objectToWorldTransform method and used that.
The way it works is you pass coordinates relative the object and the method gives you the corresponding coordinates relative to the world.
Fer instance-
offsetX = me.model.ObjectToWorldTransform(new Vector3D(0,0,0)).X
OffsetX is a double precision number. me is the star object, model is the star’s model object. Calling the method I pass in the coordinates for dead center, then have it give me just the result stored in X. I could, quite honestly, do X and Y in the same go using the vector3ds directly, but this works and I’m lazy.
On a side note, I’m splitting these posts off into their own category, under the 4x category which is under programming.
So, I tested my map routine with a second star. The star positions and everything just fine, but the label doesn’t. Have to work on that.
I’ve been programming a crude clone of Master of Orion. I mean, sure, I’ll be adding my own spin on it, but at the core of its being it is simply that.
I’ve gotten pretty far in a short amount of time. The map structure is laid out, all in XML so it’s easy to edit and apply to other data parsing tools. Mechanisms to lay stars out in a visual map, and give them name labels, are also in place and working (knock on wood) wonderfully. A lot of underlying data structures are roughed out. Oh, and a routine to figure out what star you’ve clicked on is in place. I might be able to do this a different way that’s more efficient, but I’m not certain yet.