Sooo, getting back to sweep, it seems like we do this for looks and either way works.
I agree that either way works well, but the question should be what works best for what purposes.
I re-read the whole thread and realize that my last post was somewhat redundant in that the point about taper moving the a.c. inboard had already been addressed quite early by Howard and maybe Keith. Anyway, I think this point is valid and makes taper relevant more than just aesthetically. I remember the old Smoothie ads referring to "windy weather" performance, and I have talked to others who feel that their more elliptically shaped wings are more stable (for Howard, 'less prone to upsets' or 'prone to milder upsets') in gusts. To me that seems to be because the elliptical wing has the furthest inboard a.c. of any normally accepted wing. Also I note that the more extremely tapered designs like the "Firecracker" are prominent at the top of the lists in Australia. Note also what elliptical and highly tapered wings have been on top in classic there.
So here are more extreme examples of useful taper. I hope that you'll forgive my using borrowed internet-posted pictures, a couple from Stunthangar members. The first picture shows a straight-tapered wing with the same a.c. position as an equivalent elliptical wing superimposed on a "Firecracker". The second is the Bond "Bandolero", with extreme taper. The third shows the 2013 Classic winners "down under": Palmer T-Bird, Rondinelli "Venus", and extreme-tapered Aldrich "Magnum." Configurations may be coincidental due to other reasons (Bob Palmer is quite popular there), but I think it shows that the accelerated-taper elliptical wing is very much at home in that element.
edit: I've added a final picture of my short-lived high-aspect-ratio model from 2007, which featured a highly tapered wing. It flew better than I expected in mild gusts for an Aspect ratio of over 6:1
I usually double taper mine in order to ensure that section thicknesses remain constant or increase just slightly spanwise for the reason stated above and for consistency.
SK