/cc @nigeltao

I think this may be the source of several hours of confusion I have had trying to work through a tutorial on OpenGL.

http://www.opengl-tutorial.org/beginners-tutorials/tutorial-3-matrices/

In the above tutorial the order of multiplication for the MVP matrices is as follows (at the bottom under "Putting it all together")

glm::mat4 mvp = Projection * View * Model; // Remember, matrix multiplication is the other way around

Using golang.org/x/mobile/exp/f32 multiplying in this order does not work, as was stated by @taezaz. But it seems that to match the order of operations in the tutorial I needed to transpose both the P and V matrices before multiplying them

    projection.Perspective(45, float32(float64(sz.WidthPx)/float64(sz.HeightPx)), 0.1, 100.0)
    view.LookAt(viewEye, viewCenter, viewUp)
    transpose4(projection)
    transpose4(view)
    mvp.Mul(projection, view)
    // ... now serializing the mvp matrix to __column major__ order and multiplying MVP*position in the shader work as I think it is supposed to 
    _mvp = [16]float32{
        mvp[0][0],
        mvp[1][0],
        mvp[2][0],
        mvp[3][0],
        mvp[0][1],
        mvp[1][1],
        mvp[2][1],
        mvp[3][1],
        mvp[0][2],
        mvp[1][2],
        mvp[2][2],
        mvp[3][2],
        mvp[0][3],
        mvp[1][3],
        mvp[2][3],
        mvp[3][3],
    }

To me this seems to suggest that both the P and V matrices are using row major (edit: column major) order? But I could be very wrong. I'm very new to OpenGL. It's also been a while since I did much linear algebra. And like I said. I'm pretty confused right now.