Thanks for the responses, guys. I was combining view into projection because on the directX end, I'd set my camera in the view matrix, then move meshes around using the world matrix. I was trying to find a way to do it in OpenGL that wouldn't force me to basically set the camera every time I place a mesh.
So what's the right way to do this...? Take my "world" matrix, multiply it by "view," and then set that as ModelView? I'll probably keep my camera matrix seperate, and then just multiply the "actual" modelview matrix every time I move a model. Does that sound like a viable solution?