No please don't shut up it's really interesting. What commands would be needed in OrthoContext (and BTW why not contribute ?)
Two approaches might be taken with ortho context. Either emulate what we see in the GL context:
OrthoContext.vertex(float x, float y);
OrhtoContext.texCoord(float u, float v);
OrthoContext.startQuad(); // or strips
OrthoContext.startTriangle(); // or strings
Or you could make it a bit more Graphics2D like with
OrthoContext.drawRect(float x, float y, float width, float height);
It'd be more work initially but it would also make Xith very easy to understand when working with GUIs. You could just draw whatever you wanted to over the top of the scene in an accelerated way. Naive implementation would be just to proxy the methods onto a GL context configured for orthographic projection - more complicated might be to cache the calls into arrays and spit them to card in batches.
If someone came along an implemented a DirectX layer you just translate the calls into whatever it needs to be for DirectX - which I seem to remember is pretty similiar at this basic level anyway.
Re: Contribute - I don't use Xith active and I have plenty of other stuff to do at the moment. I'm also not too keen on the style of responses I've seen with respect to contributions recently.