For a few years in the 70s we taught our introductory program starting
with *two* two-unit courses in parallel. In one course we taught
conventional mechanics, starting with kinematics. In the other course
we started with optics and finished with special relativity (out of
Taylor and Wheeler "Spacetime Physics"). I thought the scheme worked
well for the reasons cited in this discussion. My colleagues didn't.
The curriculum was abandoned because it was too different, and because
one could not use a single textbook in serial order.
The mix was chosen for its low level of mathematical sophistication.
The course was meant to start after one semester of calculus, and
optics and relativity can mostly be done without any calculus (only
Fermat's principle and some physical optics requires calculus).