- From: Simon Fraser <smfr@me.com>
- Date: Tue, 13 Jul 2010 10:18:22 -0700
- To: Brad Kemper <brad.kemper@gmail.com>
- Cc: www-style list <www-style@w3.org>, fantasai <fantasai.lists@inkedblade.net>, Brian Manthos <brianman@microsoft.com>
On Jul 13, 2010, at 9:11 AM, Brad Kemper wrote: > On Jul 13, 2010, at 8:58 AM, Simon Fraser <smfr@me.com> wrote: > >> On Jul 13, 2010, at 8:52 AM, Brad Kemper wrote: >> >>> On Jul 13, 2010, at 8:36 AM, Simon Fraser wrote: >>> >>>> Surely this isn't an argument in favor of this behavior for blur? The "outward blur only" behavior does not have to have any special discussion for the handling of odd values for the blur amount, but "entire blur region" does. >>>> >>> >>> It does seem to me that an odd number of blur pixels should be possible. >> >> I think you'd be hard-pressed to tell the difference between a 1.5px and 2px shadow, > > Depends on the resolution of the device. On large, low res device (electronic scoreboard, for example), I think I could tell the difference between 3 pixels and 4 pixels. > > For animations and transitions of different shadow blur widths, I'd want it to be as smooth as possible, adding or removing one pixel of blur at a time. > > And when 1px = 2x2 or more device pixels, I don't see any reason not to allow fractional CSS pixels, especially to make animations smoother. The spec allows for fractional pixels (it's a <length>), so animations/transitions can interpolate smoothly either way. And the fact that fractional pixels are allowed means that this "odd pixel" issue doesn't weigh in on either side of the argument. Simon
Received on Tuesday, 13 July 2010 17:18:56 UTC