Of course, otherwise it would be way too hard to dodge every hostile thing coming at you.

Textures will come later (they'll come ), but yeah I forgot about clipping I'll take care of that, even if it's not a functionality.

Well, as everyone in this thread said before me (), if you can move one bullet, you can move 100 of them with a For loop. This is where arrays are useful.

The things you call "arrays" don't really exist in Axe, since there aren't any data type nor anything. What's there is plain RAM. But you can easily use it as an array-ish variable.

For example, let's use L₁ ([2nd][1]). It's a pointer on a 768 bytes free RAM area, so it should be enough. Although it's in fact a simple area that you can go through using {r₁+L₁} where r₁ goes from 0 to 767, you can also use another technique : use {r₂*X+r₁+L₁}, where r₂ goes from 0 to the number of rows minus 1 of your "array", X is the number of colons in your array and r₁ goes from 0 to X - 1. Thus, you can pass r₁ as the colon of an element of your array, and r₂ as the row. And poof, you have an array.

Isn't the squared distance (X-Xyou)^2+(Y-Yyou)^2 ? I actually use that in IkarugaX, it's pretty fast and accurate.

I see what I think is an optimisation, I'll just make an Axe-like algorithm out of it :

Lbl GSortID
...
GDB1ZFACE is a nbFaces*2 large free area
GDB1FACE contains the face definition in the form of Data(0,4,2,...)
GDB1ZVERT contains the 2-bytes Z coordinate of each vertex
GDB1SORT is a nbFaces large free area
...

.Z-average value for each face
For(r₂,0,nbFaces)
  0→{r₂*2+GDB1ZFACE→r₃}^r
  For(r₁,0,nbVerticesForThisFace)
    {{r₂*nbVerticesForThisFace+GDB1FACE}*2+GDB1ZVERT}^r+{r₃}^r→{r₃}^r
  End
  {r₃}^r//nbVerticesForThisFace→{r₃}^r
End

.Sort
While 1
  0→F
  For(r₁,0,nbFaces-1)
    If {{r₁+GDB1SORT}*2+GDB1ZFACE}^r<<{{r₁+GDB1SORT+1}*2+GDB1ZFACE}^r
      1→F
      Exch(r₁+GDB1SORT,r₁+GDB1SORT+1,1)
    End
  End
End!If F
Return

With that you have your list in GDB1SORT. I can't really test it since I'm really not familiar with GLib's environment, so I let you test. Ask if you don't understand something.

Then don't create any ID list, but directly sort the distance list instead. So you'll come up with a single sorted list.

I'll try to implement the sorting method I used with AxeJh3D.

So in GSortID, you build a list of X^2+Y^2, and then sort it from greatest to lowest ?

Also, these speed tests are made at 15MHz obviously ?