Brad Kemper wrote: > On Oct 1, 2009, at 11:59 AM, "Tab Atkins Jr." <jackalmage@gmail.com> wrote: > >> I really don't have a problem with the UA deciding whether to upscale >> or downscale here, since it's a fairly small difference anyway. But I >> agree that the upscaling problems are pretty small, so it wouldn't be >> a bad thing to use a strict algorithm that treated all images equally. > > Yeah, my view is that the distortion problems when doing large > downscaling are worse that the resolution problems when doing small to > medium upscaling. So even if the algorithm didn't favor downscaling, the > upscaling would not be bad enough to worry about. Been thinking about this, and I think you're right. For border-image and double-constrained background images, this would cause significantly more distortion because the other dimension is fixed at this point. If I was better with graphics I'd draw a picture to illustrate this. :) But imagining a row of stars 5, the last one of which only 1/3 fits. If the height wasn't constrained, we could just shrink the stars so you'd get 5 small stars, which is still ok. But that's not the case. We'd have to shrink the stars a lot width-wise, but height-wise they can't change. So they'd turn into 5 rather anorexic stars, rather than only 4 slightly overweight ones. So I'm leaning towards just straight-up rounding at this point. Bert, comments? ~fantasaiReceived on Saturday, 3 October 2009 05:44:50 UTC
This archive was generated by hypermail 2.4.0 : Monday, 23 January 2023 02:13:40 UTC