- From: Alan Gresley <alan@css-class.com>
- Date: Tue, 20 May 2008 19:52:37 +1000
- To: David Woolley <forums@david-woolley.me.uk>
- CC: "www-style@w3.org" <www-style@w3.org>
David Woolley wrote: > > Alan Gresley wrote: >> >> >> Boris, what websites depend on a behavior that is clearly undefined? > > Most? Most sites are designed empirically, based on the behaviour of a > particular browser (or are cut and pasted from such sites) rather than > from an understanding of what the specification actually says. Sites > rely on undefined behaviour or are reckless as to whether or not it is > undefined. > > That's precisely why there is so much demand in the web world for all > "invalid" constructs to have precisely defined behaviour (or as I would > see it, for there to be no invalid or undefined constructs). If an author chooses to have a base layout such as: <div style="float:left"> <div style="float: left"> <div style="overflow: hidden; margin-left: 200px"> <div style="float:left"> <div> .... </div> </div> </div> Then they will need to learn how to send different style rules to every major version of every implementation such as: Gecko 1.7 Gecko 1.8 Gecko 1.9 Opera 8.50 Opera 9.10 Opera 9.50 Safari 3 IE6 IE7 IE8 It's possible to do such a thing but why should this have to be done in the first place? Alan
Received on Tuesday, 20 May 2008 09:53:36 UTC