RE: Special fonts don't matter for font-family change.

Wayne wrote “We are only verifying testability, not whether it will look good once an AT grooms it.”

This seems to miss the point for me – if the font change causes icon fonts to turn into squares there is loss of content for me.  IMO this SC has nothing to do with AT grooming whatever that might be.

Jonathan


From: Wayne Dick [mailto:wayneedick@gmail.com]
Sent: Wednesday, April 05, 2017 12:57 PM
To: GLWAI Guidelines WG org; public-low-vision-a11y-tf
Subject: Special fonts don't matter for font-family change.

We are only verifying testability, not whether it will look good once an AT grooms it.
So, suppose an author uses an icon font, or a math font, or a script font for emphasis or scientific semantic reasons, it does not matter if that font is changed when testing if CSS can change the font family. If the special font is changed to tahoma or verdana or times new roman, it is unimportant. Unassigned character codes  are replaced by a box containing the character code as an integer.  That is ok, it changed to the target font. The display image isn't pretty, but the fact that the character changed is demonstrated as true.
All you need is  <style> * {font-family: "Whatever" !important;} </style> inserted in the last position of the <head> element and the test will give you a true or false.
What will it measure?
The answer is this. Did the author get in the way of CSS change to font.
If all text does not change to the "Whatever" font then the author has done something to override CSS in an inaccessible
At present, semantic markers for font usage are not extensive enough, and WCAG 2.0  did not recognize inappropriate use of font to convey meaning as a violation of 1.3.1. So, AT will have to do what screen readers do when faced with semantically unclear situations. It will have to make a good guess as to why the author changed the font family mid line.
Wayne

 way.

Received on Thursday, 6 April 2017 13:04:45 UTC