- From: Shane McCarron <shane@aptest.com>
- Date: Tue, 9 Feb 2016 08:53:18 -0600
- To: Ivan Herman <ivan@w3.org>
- Cc: Janina Sajka <janina@rednote.net>, Florian Rivoal <florian@rivoal.net>, James Craig <jcraig@apple.com>, W3C PF - DPUB Joint Task Force <public-dpub-aria@w3.org>
- Message-ID: <CAOk_reEJo5dsH1bQ4V=L3xbLizhf169ndJebVqR9C8T6OgPqtg@mail.gmail.com>
Ivan, A couple of comments: On Tue, Feb 9, 2016 at 6:36 AM, Ivan Herman <ivan@w3.org> wrote: > > Obviously, this is a much more general and powerful direction for CSS, and > I do not know whether it is realistic to expect anything in this direction. > (I admit that I did not even think about this until, on the call, I heard > that such discussions have already taken place before). But, let me dream: > having the browsers providing a set of user interface 'flags' that can be > queried from CSS via media queries sounds like a very useful general > feature to me... > I agree that, in general, it would be useful to be able to query user agent settings. However, it is unlikely that there is going to be a standard set of such settings. I have this nightmare scenario in my head where there are 2 that the W3C can agree upon, and then there are 1000 that start with edge- or moz- and people start using them all over the place and then we are stuck supporting them forever. We've been down that road before. Hell - we are *still* on that road! > > James, you ask "Couldn't that just be an AT behavior?": I am not sure it > is the same (and, consequently, one may not exclude the other). What I > believe that would mean is that, by default, the content of the <details> > element would be, say, invisible as set by the corresponding stylesheet, > but the AT can overrule this and make that content accessible (eg, as > text). But is it a good idea to hardwire such behavior into the AT? Isn't > that something that the end user could somehow control even if he/she does > not use a particular AT, just an average browser instead? It, sort of, > feels at another level to me. > Actually, forgive me, but this isn't an AT feature at all. Or rather, it isn't exclusively an AT feature. From a high level, this is very much a "some users want to be exposed to different information than other users". This dovetails with work in the Cognitive group, but again, that is not exclusively an A11Y problem either. Consider the HTML specification. That document has / used to have a button that said "show me developer details". If I am an implementor of a user agent, I want that enabled. Probably all the time. And I don't want to have to select it every time. In fact, I really want it enabled all the time for all specifications. If the W3C had some standard for that, it would be really handy. The general case is just that W3C microcosm writ large. Can MQ solve this problem? Sure. Probably. Let's really brain storm about this though. Could we solve the general case problem? In an in-the-wild-extensible manner? Maybe in conjunction with user agent preferences or general case user preference storage external to a user agent. Imagine a MQ that translates as "user's preference for X is true or false". Where X is scoped to organizational vocabularies. @media( -userpref-wai.w3.org-high-contrast ) { } Then, in user agents that support it, the user agent requests/enables access to the user's custom preferences. These are local, or in the cloud, or whatever. Now that's sexy. Combine it with something like enquire.js and you have a super powerful general case user preference mechanism that can be used statically and dynamically. No way to actually SET the preferences, mind you. But it feels like something that would be easily supported through extensions if something like this actually went anywhere. But I digress. What is the basic problem we are trying to solve here? Some users want the extended descriptions of things available all the time, and some do not. And the mechanism for querying that user preference should be consistent so that content authors can take advantage of it in a well-defined manner. I know that a lot of us have spent a ridiculous amount of time trying to resolve this seemingly-simple problem. There are a couple of proposals on the table right now. This very basic MQ proposal seems like one that gets us over the hump right now. If this proposal isn't feasible or palatable, can the objectors please suggest alternative(s) with examples? That would be a big help! P.S. If anyone wants to work on a general user preference setting / getting mechanism with me (tacked onto media-queries or not), I am up for it. -- Shane McCarron Managing Director, Applied Testing and Technology, Inc.
Received on Tuesday, 9 February 2016 14:53:54 UTC