- From: Mihai Sucan <mihai.sucan@gmail.com>
- Date: Thu, 31 Aug 2006 22:53:12 +0300
Le Thu, 31 Aug 2006 22:25:45 +0300, Simon Pieters <zcorpan at hotmail.com> a ?crit: > Hi, > > From: "Mihai Sucan" <mihai.sucan at gmail.com> >> I don't think we want another attribute. I tend to like the idea of >> having tabindex="-2". IIRC -1 is already "reserved" for skipping the >> control when navigating (maybe only in some proposals?). > > All negative values are for taking the element out of tab order. But why > can't all values mean that the element is a scoping element? Perhaps > someone wants the scoping element to be in the tab order aswell (can't > think of a particular use-case off the top of my head though): > > <div tabindex="2">4 > <span tabindex="2">6</span> > <span tabindex="1">5</span> > </div> > <div tabindex="1">1 > <span tabindex="2">3</span> > <span tabindex="1">2</span> > </div> Yes, good idea. One also needs to define the tab order of the "groups". But tables, divs, spans are not normally focusable. >> I can agree with tabindex="scoped" - it's good enough too. > > But how do you deal with the .tabIndex DOM attribute? It can't be long > and DOMString at the same time. Changing it to DOMString might break > some sites. Currently, if you for instance do: > > document.body.tabIndex = "scoped"; > > ...then it results in: > > <body tabindex="0"> True. But I don't believe many sites will break because of this. Remember many web developers don't really care about types. They don't even know much about them. IMHO the number of sites that might break makes these cases neglectable. -- http://www.robodesign.ro ROBO Design - We bring you the future
Received on Thursday, 31 August 2006 12:53:12 UTC