The glorified T-buffer is just an accumulation buffer (Something supported in OpenGL since version 1.0). It's just that due to the way the voodoos were designed, the multi-chip versions could have several buffers (One per chip, actually), and would then combine them all quickly. Getting effects like anti-aliasing, motion blur and depth-of-field is just simple blending between frames.
For anti-aliasing, you draw frames with a slightly different view-point and blend them.
For motion blur, you simply draw X times faster (Say, four times as fast) to the off-screen buffers, and then blend every X frames and display the result on screen (The technique is called 'Temporal Anti-aliasing').
For depth of field, you draw several frames with different FOVs, with the FOVs differing according to where you want the focus to be.
It's nothing at all compared to shaders. Everything done with T-buffers can be done without them easily.
Edit: Lesson for life - Don't get me started on graphics
