- From: Max Romantschuk <max@provico.fi>
- Date: Tue, 11 May 2004 09:21:21 +0300
- Cc: www-style@w3.org
Richard York wrote: > gradient([[<color>{1,}] <effect>]) > > Whereas grandient would be a color type itself and be a value of any > property taking a color value. > > color: gradient(black white ltr); > > A little OT, but in the same neighborhood.. this would be a cool > possibility too: > color: url(picture.jpg); *snip* > I don't quite see the argument for gradients not belonging in CSS, you > could argue the same thing about opacity, just use PNG instead. I think > gradients open up design possibilities, and I thought that's what CSS > was all about! Personally I like any idea that eliminates all the fuss > of dealing with images. One (wild) idea is to introduce a texture concept into CSS. Currently CSS deals with colors and backgrounds. Images are always a special case. If a texture concept was introduced it might make things more simple in the end. A texture could basically be anything, and the whole color/gradient/image/tiled-image deal could be wrapped up in the texture. Then we'd end up with a two-stage process: 1. Define textures. 2. Apply testures to backgrounds, borders, text etc. This would be rather different from the current implementation, but it would allow for great flexibility. A UA could basically render a texture to an internal bit map, and then apply it to the target surface. An additional benefit would be the potential for hardware acceleration, although this is just a side effect. With textures there would be less to keep beefing up the background and border properties, as most additions could be done to the textures themselves. How do other people feel about this idea? I realize it's radically different from the current implementation, which means backwards compatibility is a major issue. Then again XHTML 2 won't be backwards compatible as far as I've understood, so that's maybe something to target? -- Max Romantschuk http://max.nma.fi/
Received on Tuesday, 11 May 2004 02:21:45 UTC