- From: Tab Atkins Jr. <jackalmage@gmail.com>
- Date: Wed, 12 Dec 2012 13:34:45 -0800
- To: Sylvain Galineau <sylvaing@microsoft.com>
- Cc: Florian Rivoal <florian@rivoal.net>, "www-style@w3.org" <www-style@w3.org>, "Edward O'Connor" <eoconnor@apple.com>
On Wed, Dec 12, 2012 at 1:25 PM, Sylvain Galineau <sylvaing@microsoft.com> wrote: > [Tab Atkins Jr.:] >> Compare this to today's world, where in the "best" case you decide whether >> to do touch or mouse stuff based on a combination of user-agent and >> device-size detection, and new devices generally get sorted semi-randomly >> into one of the categories. > > Yes, they get sorted into a category but they remain in that category, right? > Are there apps out there that will change their layout on me based on when > they think I'm using one pointer or the other? Maybe I need to look at them > to understand why anyone would want to risk that. Oh, I wasn't thinking that it would swap. I mean, maybe it could, but I was assuming that the MQ would remain stable. Though, that seems somewhat bad if you have a device that is primarily mouse-driven, but you can disconnect the monitor to use as a touch tablet. You'd presumably want the page to change slightly to accommodate that? I don't think it's really possible to *tell* whether you're using one type of pointer or another today, so you won't find any apps that deal with that information. >> With this, though it may not be able to give the best answer every time, >> it most circumstances it can (as there's only a single interface method). > > Not sure what 'it' refers to here? it = the MQ, except for the second "it", which is a typo for "in". >> Maybe we can extend the feature to handle mixed-interface devices better? >> I'm not sure how to, off the top of my head. > > All right, let's assume you can detect that. What CSS do you write when you > know the UA has both fine and coarse pointers? More specifically, what CSS > would you write that is different from that which you'd apply if it only had > one of them? I think that, ideally, I'd like to know what the user is doing *right now*, and lay out accordingly by growing/shrinking some things. In the absence of this kind of magical information, I suspect I'd be pessimistic and just always design for the touch case. Slightly larger touch targets aren't *bad* for mouse-based interaction, they just produce a slightly larger interface than is strictly necessary for a mouse. Given that, I think my answer to the "multi-model device" question is actually that it should just be pessimistic and answer "coarse". I guess that applies equally to the "hover" MQ - it should return false if the device can reasonably be used in a hover-capable and non-hover-capable state. If the device had a strong signal one way or the other, I could see it being useful to go ahead and be accurate, but I'd leave that up to the UA. ~TJ
Received on Wednesday, 12 December 2012 21:35:38 UTC