Re: How do we distinguish between I, l and 1 and 0 and O?

> How long will users be able to change font?
As long as browsers allow extensions that manipulate the user’s view. I have seen no sign of that changing.
Using “!important” in CSS can be over-ridden, the only things preventing changing fonts I’ve found are images of text (already covered), and iframes (embedded pages). That might not be a blocker, I haven’t looked into it yet.
Is there anything else you have come across? Please ping me examples if you have.

> What about when encapsulated web components come in.
That will probably improve the situation because they will be part of the main DOM, so easier to manipulate in the browser. That is a guess at the moment, I haven’t tested it.

> If someone puts an icon is a place where it can be misinterpreted as text then it needs to be identified semantically.
I agree, but what is the standard for that? As we discussed in the call, a violation of 1.3.1 needs to be identified against the (HTML) spec.
For example, heading use is outlined here:
https://www.w3.org/TR/html5/sections.html#the-h1,-h2,-h3,-h4,-h5,-and-h6-elements

Maybe we can raise a bug against HTML, either for images:
https://www.w3.org/TR/html5/embedded-content-0.html#alt

or common idioms:
https://www.w3.org/TR/html5/common-idioms.html#common-idioms


> Font substitution is easily testable and with reasonable semantic support, i.e. wrap icon fonts in <abbr title="icon meaning">icon</abbr>.
That could be a recommended approach put in the bug, but until that approach is visible to the world there is no agreed standard for authors or user agents. Therefore nothing for WCAG to hook into for conformance.

> I do not see why we should drop font family at all.
We should not “drop it”, we should make sure we enable it. However, that is not the same thing as having a bullet point in SC text.

Cheers,
-Alastair