UA/platform issues (was Length of line)

Patrick wrote:
> The disconnect for me comes when UAs/platforms that don't provide that particular functionality/mechanism are then also excluded...so effectively we're saying "a mechanism for X should be available, which in most cases is already taken care of by the UA, except this SC doesn't apply if the UA doesn't provide this sort of mechanism"?

There are different variations on this between SCs so I’m going to expand slightly:

Line-length: We are retiring because it can be achieved by the user if the reflow SC is accepted.

Reflow: This can work for HMTL (including mobile) and PDF (assuming you consider the reflow function to be acceptable on PDF readers).
There are things authors can do that torpedo a user trying to apply their own linearization to do with reading order and layout methods. I have a little proof of concept bookmarklet [1] that removes the layout for *visible* items that works really well on the test page, but with a wild variety of results across sites. Some work well, some well enough, some become unusable.

There is work to do to better define what the details are in terms of techniques & failures, but I think the concept is strong and there is at least one project outside WCAG working on a browser extension specifically for this.

Font Family: Works for HTML and PDF. For websites it is possible to change font-family now, but there are things the authors can do which ruin that attempt. Again, I think there is more work to do defining what works and doesn’t, but the requirement is worthwhile and it appears feasible.

Spacing: Works for HTML, does not work for PDF. That makes it an interested case, as it has the same sort of impact on users as font-family, but do we have to drop it because it doesn’t work with PDF? Does having this SC help move people towards a more (easily) accessible technology? Do we add an exception for non-HTML technologies?

Kind regards,

-Alastair

1] https://alastairc.ac/tests/layouts/percentages.html (the links on the right for ‘linearise’)