- From: Tab Atkins Jr. <jackalmage@gmail.com>
- Date: Thu, 1 Oct 2009 13:59:48 -0500
- To: Brad Kemper <brad.kemper@gmail.com>
- Cc: Mikko Rantalainen <mikko.rantalainen@peda.net>, www-style list <www-style@w3.org>
On Thu, Oct 1, 2009 at 1:03 PM, Brad Kemper <brad.kemper@gmail.com> wrote: > > On Oct 1, 2009, at 12:10 AM, Mikko Rantalainen wrote: > >> As you say, it makes sense to allow this for "high" resolution source >> images. I think that the spec cannot specify reasonable exact limit for >> changing from real rounding to ceil() method without knowing the scaling >> algorithm. > > I would much rather see interoperability in terms of how many tiles are > shown. > > I don't personally think the UA should be deciding based on resolution if it > should widen the image or not. It does not do that for the the 'scale' > keyword, where I can scale a raster image 1000 times bigger if I want. It > also does not do that for border-image-width, where I can turn a 5px tall > image slice into a border side that is 500px thick. > > The upscaling for round is really fine, even on auto sized raster images, in > my view. Here is a test I did: > > http://www.bradclicks.com/cssplay/border-image/round-test1.html > > I think if the author didn't want upscaling with 'round' they would create > the image tiles at a size that prevented it from happening. 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. ~TJ
Received on Thursday, 1 October 2009 19:00:42 UTC