Rending fluid using per face normals instead of smoothed normals is trivial part of otherwise quite complex task.
That makes a lot of sense, I realize now that if anything calculating all those surface normals for flat shading could make it even more complicated. But wouldn't you say that the resolution (or how many particles are included) would have a huge impact? The difference between realistic looking water and this could potentially be a huge drop in number of particles involved when creating the mesh.
Fluid physics is something that is usually extremely expensive to compute. It's basically a particle simulation where the particles interact with each other (uh-oh). Rendering is then done by generating a mesh from these particles. It'll probably work for small bodies of water, but the code needed for it will be quite difficult to code and require shaders for rendering, at least if you want anything close to realistic water behavior. Your best bet is probably to look up papers on fluid dynamics and try to implement one of them.
Thank you for being honest about fluid physics memory usage. Might make for it not being a realistic element of the game...
Also, your explanation got me on the right track as to what is going on during fluid physics. I'll have to read more into the math behind fluid dynamics.
Also, I found this really in depth 17 part article on fluids in games so hopefully it will help some.
http://software.intel.com/en-us/articles/fluid-simulation-for-video-games-part-1/