On Fri, May 28, 2010 at 4:52 PM, Robin Berjon <[email protected]> wrote: > Hi Jim, > > your comments reach us right after the WG decided to take the specification > to CR, but thankfully I was a bit slow with the editing so that we could take > them into account :) > > On May 27, 2010, at 22:42 , Jim Allan wrote: >> View-mode: fullscreen. It is not clear whether fullscreen includes a full >> set of chrome, or includes no chrome. You mention 'chrome' in the >> 'windowed' and 'floating' viewmodes. For consistency, chrome presence should >> be noted in fullscreen. > > That's correct, I've now clarified this by adding a mention of chrome for > both fullscreen and maximized. > >> It should be noted that the User Agent Accessibility Guidelines 2.0 [1] has >> success criteria that allow the user to override author settings for a >> variety of viewport view-modes including the inclusion/exclusion of >> 'chrome.' > > Yes, and that's fine. The idea here is that the UA would make a best effort > at matching the intent in a way that makes sense rather than be ultra strict. > For instance, if the app goes fullscreen but keeps a teeny bit of chrome (at > user option or not) to make it easier to exit fullscreen, then matching the > view-mode: fullscreen media query is quite clearly the right thing to do. > >> Please consider including a statement such as >> "The user agent *must* display the view-modes in a manner that meets the >> accessibility guidelines of UAAG20. " > > As much as I'd like more UAs to support UAAG I don't think that this > requirement is appropriate here. The VM specification defines a technology > with a single purpose: "if the window in which the content is being rendered > is like this, then apply these CSS style rules". It does *not* define how a > UA ought to display an actual set of window states, it doesn't in fact even > require UAs to support all the view modes. I'd expect that an application > running on an iPhone would only support maximized and fullscreen — if it > applied different style rules for each, it would still be 100% conformant. >
FWIW, I agree with Robin here. -- Marcos Caceres Opera Software ASA, http://www.opera.com/ http://datadriven.com.au
