- From: Orion Adrian <orion.adrian@gmail.com>
- Date: Sat, 4 Dec 2004 11:08:12 -0500
- To: www-style@w3.org
I do believe we've seen the limits of CSS. It was designed so that properties did not have context or structure. CSS properties often represent a multitude of actual presentation properties (e.g. float). No generic system was put in place to handle a) array properties b) combinitorial fallback (i.e. when one property fails, other properties should also fail because of their interactions) c) unambiguous identification of values in combination properties (e.g. background) d) user-defined constants A and b I think are especially destructive to long-term success of the system, though CSS's lack of context producing structure I believe will be it's ultimate downfall. This lack of context producing structure forces systems to implement and array of complex interaction maps (i.e. property x allows values y, z and w only when property a's value is b). Context would allow much simpler interaction maps at a gramatical level without introducing them as a runtime evaluation. Though the biggest gains would be in the realm of CSS authors. I am a certified master CSS author and I find keeping track of what properties affect what to be extremely difficult. The interactions of various CSS properties are convoluted enough to make you want to pull your hair out. I can assure you that this is not a simple system at all having tried to explain it to dozens of people. Simplification of grammar is a laudible goal, but not at the expense of learnability, the ability to retain it and usability. I am not complaining about CSS simply to get my own proposal accross, but rather to suggest that CSS 3 do what XHTML 2 did and break away from mistakes of the past. I am confident that a new grammer could be put together that handles all of these issues well. I am also confident that a cleansing fire can be put to many of the properties of CSS so they can be replaced by simpler, mutually-exclusive properties. Doing so I believe will promote growth in the CSS community. Letting the diseased parts of CSS survive will just make the whole forest sick. Orion Adrian
Received on Saturday, 4 December 2004 16:08:51 UTC