- From: Emrah BASKAYA <emrahbaskaya@hesido.com>
- Date: Tue, 09 May 2006 22:19:11 +0300
- To: "www-style.w3.org" <www-style@w3.org>
The things mentioned in this thread for transition effects are very strictly a scripting job. Drawing a line between basic required behaviour for accessibility (:hover), and eye-candy effects (opacity fade on hover) is actually very easy. I do understand the problem of having to put style related information right in the javascript, but these things are too heavy for CSS, IMHO: just because we don't want to update two files (css+javascript) instead of one shouldn't mean CSS should hold the burden of these behaviours. However, I do understand the problem of having to declare style in Javascript. What we actually need is standardized methods for reading CSS values based on classes/id's right in Javascript using DOM methods, e.g. a mock-up dom method for reading a class value myElementsBorderWidth = document.activeStyle('screen').getClassPropertyValue('myclass','border-width') Than with this standardized information, any transition effect can easily be programmed, the possibilities are endless. Thinking of no scriptable environments, following philosophy is just fine: .noscripting a:hover { ... } Where the .noscripting will be removed by our 'behaviour layer' when first executed from our,e.g. body element, again, a little task easily handled by the programmer. The latest in stylesheet reading is simply not anything near practical, see it at. http://developer.mozilla.org/en/docs/DOM:stylesheet In conclusion, I don't think we need anything more than :hover like simple behaviours for style sheets. They need to be there for accessibility reasons, for letting user know he is doing something with the element being hovered/focused in the most basic sense. However, anything one step more complex than that should better be done by scripting. That said, the need to read styles from classes using DOM methods is something most badly needed. I will later be working on some methods for animation using a method I'll call "class morphing", which will probably require standards abiding browsers. I just need some browser inconsistencies fixed before I can share it with the crowd (like Opera sadly returning offset sizes instead of declared style sizes when getComputedStyle is being used: http://www.hesido.com/test/webdesign/getwidth.htm ). -- Emrah BASKAYA www.hesido.com
Received on Tuesday, 9 May 2006 19:26:08 UTC