- From: Jim Ley <jim@jibbering.com>
- Date: Wed, 5 Dec 2001 23:55:57 -0000
- To: <w3c-wai-ig@w3.org>
"David Woolley" > > It should be: > > > > // once per JS file > > var d=document; > > var DOM=(d.getElementById)?true:false; > > I'd need to check some details, but I believe that you can define > functions dynamically, by the use of strings, or possibly by > copying a function between attributes. If that is correct, I'd say > the best way of doing this would be to upgrade all the browsers to > DOM1 status by defining getElementById in those that don't already > support it. If you look at GetRef in the clj FAQ http://jibbering.com/faq/#FAQ4_15 you'll see my approach (the expando property on the document isn't necessary standard so I'd consider not doing that, I could have a browser which doesn't support it for example, I also don't agree with the sensibleness of duplicating gEBI in old browsers, they differ in too many and unknown ways (and the amount that spoof/paritially implement things like document.all make it more complicated.) I think it's safer to think in DOM1 or 2 terms and give everyone else fallback it's a lot safer. "Dr Clue" a poster in comp.lang.javascript and a talented scripter has a very good "library" system that you're suggesting but it's far from a quality I could suggest being reasonable from an accessibility P.O.V. > Only resort to thinking in terms of a browser when it behaves as > though a feature is present when it is actually broken; but, better, > avoid such features. Detect the break aswell if possible, as you'll never know when it was fixed... Jim.
Received on Wednesday, 5 December 2001 18:55:52 UTC