The current target for panning heuristics seems to approximately match the motion of my finger --- if I fling my finger up X pixels, the page seems to also fling up somewhere around that distance as well. That feels frustrating physically because the actual friction of my finger on the phone touch screen is fairly low and the surfaces seem smooth (I can't see what's behind the page), but the page reacts as if I'd flung sandpaper over sandpaper.
We could be really formal about this and decide on a physical model for the page (mass and friction coefficient), then measure user's real interaction of finger with touch screen and adjust panning physics wrt that.
But maybe more importantly, feel for b2g has diverged considerably from firefox-android, with not much rationale that I've seen.
We tried to align more closely with another popular platform, which overall I find a lot more pleasing. We can reduce the friction and see what you think. Is that sufficient?
I agree that we should reduce the friction a little bit. I think the Fennec friction is too much though, especially considering the form factor of Otoro vs. most phones that Fennec runs on.
Doug can you cook up something? I can review tomorrow.
Created attachment 653472 [details] [diff] [review]
Lower async panning friction
Behold my magnum opus of patches.
Not sure who wants to review this, I just put :gal based on comment 3.