Re: Query regarding 1.3.4

On 11/8/2022 11:48:23 AM, David Woolley <forums@david-woolley.me.uk> wrote:
On 08/11/2022 17:40, Brooks Newton wrote:
> the success criterion was all about not forcing users to have to
> be able to change orientation ever to get content or functionality to work
> correctly.

As I understand it the question is not about the site forcing a change
of orientation, but the user forcing a change. I'm working on the
assumption that the site does load, initially, in both orientations.

Under benefits, the more detailed information, two examples are given of
benefits. One is that the user lacks the physical ability to change the
orientation, and the other is that they find font sizes easier to cope
with in landscape. In the first case there is no possibility of the
user trying to force a change. I guess you could suggest that in the
second, the user entered the site in portrait and decided they had to go
to landscape to be able to read it, but I'd argue that someone in that
situation would always try to use landscape.

Try framing this same question through the lens of the Reflow criteria. Reflow requires testing pages at a 320px/256px viewport - but the intention of the SC was not for mobile users on small screens. The intention of that SC was for users who browse with significant magnification, and it requires that users do not need to scroll in more than 1 direction to view content, all other SC are met when in reflow, and no functionality is removed from the page. Elements can transform into different structures (footer links presented as a list on desktop but in a disclosure on smaller viewports) but they still have to be available.

Brooks' example of a person with a caregiver who doesn't return the device to the orientation that the user needs is a good one, but just ONE of the many reasons why someone may open a page in one orientation and then need to switch to a different orientation.  I can imagine a dozen more, and they are not edge cases.

After re-reading the SC and the understanding doco, then reading through this GitHub issue [https://github.com/w3c/wcag/issues/391], I am less convinced that WCAG actually requires opening the page in one orientation and then changing the orientation and testing the page, though I believe this is what should be required and perhaps what the WCAG folks thought they were saying, but actually didn't.