Yes; We'll see...
...but, before I shut up, something you just wrote compells me to say something about supersampling:
It is not just a matter of reducing jaggies on edges; Just like the word states, you are taking more samples from the world, picking denser: More detail goes, averaged, into each rendered pixel, and it is closer to geometrically correct; Not just for polygon edges, but also with textures; More so than you get out of the mipmapping, becase mipmapping is prequantised, and doesn't care if you've moved your head a millimetre to the side - your current mipmap is what it is, and attains a "picture-like" sense, more than a "world" one.
There is a quality that I find hard to describe. when you have a healthy amount of supersampling, with our low-ish resolutions: The virtual world feels more "solid", "stable", "real", and it is not only because there is less pixels popping in and out of existence, and crawling along polygon edges, because hand in hand with those things, you get everything not "jumping" from pixel to pixel, when you pan across the landscape, but moving smoothly instead -- this, amazingly, even when dropping frames.
Now; If one think of the imagery as a brook of flowing water, the non-integer-scale-factor upscaling should to a degree come across as a grid of "standing waves", that displaces the image, like the way a rock on the bed of the brook makes the moving water "bulge up" as it flows over it. This would disrupt that nice feeling of stabilty, that supersampling afforded us, producing a "shimmering" quality, kind of like projecting onto coarse fabric and moving the projector around. This is regardless of how much one supersamples, because it occurs after the entire supersampling procedure is performed and done with -- better than with no supersampling, but then; Everything is. :7
Will love it, when I'm proven wrong. :7