zoebijl / font-awesome Goto Github PK

What combinations should we test? We don't want to have too many different combinations across the various tests.

Something like:

JAWS with Internet Explorer on Windows 8+
JAWS with Firefox on Windows 8+
JAWS with Chrome on Windows 8+
NVDA with Internet Explorer on Windows 8+
NVDA with Firefox on Windows 8+
NVDA with Chrome on Windows 8+
Orca with Firefox on Ubuntu 12.10+
VoiceOver with Safari on OS X 10.10+
VoiceOver with Safari on iOS 9+

The row headings only read the AT name, so if you have VoiceOver on both OS X and iOS it will just read “VoiceOver” twice. Might want to include browser combo in first row?

Moar!

We need this for test two: visually hidden text.

See Karl Groves' fiddle with Hans Hillen's code.

We need to investigate what exactly is tripping up IE/JAWS. Is it that image role isn't seen as valid enough to read out an aria-label, or is the problem aria-label itself, or does it have to do with how IE deals with characters/glyphs it doesn't normally know?

From this PR: #16

To have meaningful results in the plain test bed we need to test more icons.

ToDo

• Add a bunch of icons to test bed
• Test bunch of icons on test bed page

Not sure what was coming over me.

Someone on the list mentioned this. Not sure if it will help, but we'll have to science it!

More backstory to the whole discussion maybe?
Put final test result in README?

We should put the output of screen readers in samp tags.

This project doesn't have enough cats.

Row headers alternate in colour, as do the rows themselves. Is there enough contrast for colour blind people?

Search for—and implement—more test examples

To fix issues not solely related to screen readers:

• Users who replace all fonts (e.g. dyslectics).
• Multi colours

Much possibilities:

• Mark text red
• Make background red
• Mark with an icon
• Mark with red border

That's just visual, need text alternative.

Best would be an image with alt text.

<img src="fail.svg" alt="Failure:"> or something.

See issue #9.

Which end-user tools and techniques related to text customization should we use for testing?

Smarter people than me have defined "text customization":
http://mitch11.blogspot.com/2012/11/my-notes-from-w3c-readability.html

Here are some examples to illustrate what I mean by "tools and techniques":

• Internet Explorer "Ignore font styles specified on web pages" -- http://windows.microsoft.com/en-us/windows/change-webpage-colors-fonts#1TC=windows-7
• Firefox "Allow pages to use their own fonts" -- https://bugzilla.mozilla.org/show_bug.cgi?id=789788
• Plug-ins -- http://superuser.com/questions/857711/change-chromes-font-for-a-specific-website

These initial examples are just for illustration. They don't necessarily reflect what people with disabilities are actually using today.

This ticket is to go find out which tools and techniques actually matter for people.