- From: Brad Kemper <brad.kemper@gmail.com>
- Date: Wed, 12 May 2010 01:33:42 -0700
- To: Brian Manthos <brianman@microsoft.com>
- Cc: Boris Zbarsky <bzbarsky@MIT.EDU>, "www-style@w3.org" <www-style@w3.org>
Received on Wednesday, 12 May 2010 08:34:19 UTC
On May 12, 2010, at 1:20 AM, Brian Manthos wrote: >>> On 5/11/10 1:49 PM, Brian Manthos wrote: >>>> For values below 0.00833333331px, Firefox continues to show no image. >>> >>> Right. Lengths in Gecko are stored as integers in units that are 1/60 of a CSS px. >>> The number above is about 1/120. So things smaller than that would would round to 0. >> Why don't things less than half a device pixel round to zero (at least for used value)? >> That is what I would expect (unless you were going to simulate a 60th of a pixel by >> averaging the colors of the subpixels, the way anti-aliasing or image size interpolation >> does). I assume you have a perfectly logical reason, but I cannot guess what it is. > > Half a device pixel before or after zooming? Before. Multiplied by zoom factor. When I zoom out in Safari, 1-pixel black lines against a white background become gray lines. > > How should it behave when printed to a 1200dpi device? Should lines "appear" that don't appear in the on-screen/monitor rendering? > Of course. If I specify a line that is 1/1200 of an inch, I'd expect a faint hairline on the paper, but not much of anything on the screen (or maybe a very faint line simulating the thinness of it). I wouldn't want it to look 13 times bigger on my 96dpi screen.
Received on Wednesday, 12 May 2010 08:34:19 UTC