- From: Ian Hickson <ian@hixie.ch>
- Date: Mon, 19 Dec 2005 22:26:25 +0000 (UTC)
On Tue, 13 Dec 2005, Dean Edwards wrote: > > Ian Hickson wrote: > > Is this only because it is easier to implement the DOM interface on a > > limited set of elements than on all elements? > > Basically, yes. As far as the DOM is concerned it is nigh on impossible > to support dynamic creation of repetition templates. Support for > repetition blocks is slightly easier. Regarding CSS/markup, the only way > to support the spec as it currently stands is to write some form of > DOM-walking to identify repetition templates and associated blocks. Ok... > I outlined a way to speed this up in a previous message: > > http://listserver.dreamhost.com/pipermail/whatwg-whatwg.org/2005-March/003299.html > > (you might have to read the entire thread to get the point). When I read that back in March I understood it as an argument for why you could implement the current spec even though it applied to all elements; but I see now that it would be faster if it was restricted. > If the tag set were restricted then the need for DOM walking would be > removed and would significantly improve performance on an IE platform. Apart from the DOM interface extensions (that apply to every element), is there any feature other than the initial crawl for repeat-start and repeat-min attributes that would be helped? -- Ian Hickson U+1047E )\._.,--....,'``. fL http://ln.hixie.ch/ U+263A /, _.. \ _\ ;`._ ,. Things that are impossible just take longer. `._.-(,_..'--(,_..'`-.;.'
Received on Monday, 19 December 2005 14:26:25 UTC