Comment on 1.4.12 User Interface Component Contrast (Minimum)

[jnurthen]

As for 1.4.11 please simplify to require a 3:1 ratio for all sizes. This will still be a huge improvement over WCAG2 where this is not required at all and will allow this to be realistically testable.

[DavidMacDonald]

Proposed response.

Thank you for your comment. We have amended the contrast ratio to 3:1. A pull request has been made. https://github.com/w3c/wcag21/pull/498

[steverep]

also graphics contrast is not as hard to interpret as text.

We should not be making such statements without backing them up with research, and such research would be incredibly complicated given the likely variance with type of graphic. This is speculative opinion, and I'll continue to argue that following the relationship of many scientific and mathematical charts is just as visually intensive as reading text, if not more so.

[allanj-uaag]

Thank you Steve. All of the researchers the Low Vision Task Force has talked to agree that the research is difficult. They also say, that at a MINIMUM graphics contrast should be equivalent to text contrast. The LVTF has done lots of research and found nothing to refute their claims. Without further evidence to the contrary anything else is opinion. Jim

On Tue, Oct 10, 2017 at 4:44 PM, Steve Repsher notifications@github.com wrote:

also graphics contrast is not as hard to interpret as text.

We should not be making such statements without backing them up with research, and such research would be incredibly complicated given the likely variance with type of graphic. This is speculative opinion, and I'll continue to argue that following the relationship of many scientific and mathematical charts is just as visually intensive as reading text, if not more so.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/w3c/wcag21/issues/490#issuecomment-335617455, or mute the thread https://github.com/notifications/unsubscribe-auth/AG_WLyQKz0jHpyO_z5fSRF67eiz4abrcks5sq-UxgaJpZM4PywOi .

-- Jim Allan, Accessibility Coordinator Texas School for the Blind and Visually Impaired 1100 W. 45th St., Austin, Texas 78756 voice 512.206.9315 fax: 512.206.9452 http://www.tsbvi.edu/ "We shape our tools and thereafter our tools shape us." McLuhan, 1964

[DavidMacDonald]

@steverep @lauracarlson @goodwitch The phrase "also graphics contrast is not as hard to interpret as text." has been removed for reasons cited.

LVTF reached out to Gordon Legge, Distinguished McKnight Professor at the University of Minnesota on this topic. Dr Legge is head of the Minnesota Laboratory for Low-Vision Research and a founding member and Scientific Co-Director of the Center for Applied and Translational Sensory Science (CATSS). He said: ...Sometimes form controls can be small, and viewed near the acuity limit. Sometimes form controls can be "crowded," that is, there are neighboring visual objects."

I agree controls can "sometimes" be "crowded" or "small". I don't think we have to do a study to establish that text is usually smaller than controls, and letters are usually closer together than controls.

I agree that higher contrast is often better for low vision. I think we have to balance that with the reality of testing and the burden on authors. And also the reality that this applies to focus rings which gives us a third layer of contrast. The button has to contrast with the focus ring which in turn has to contrast with the background. Is there any combination that can do that. Will marketing departments accept such a limited palette?

[steverep]

I agree controls can "sometimes" be "crowded" or "small". I don't think we have to do a study to establish that text is usually smaller than controls, and letters are usually closer together than controls.

That's just one statement and this response misses the point. It's not about size but about thickness. The lines that make up radio buttons, check boxes, and text entry fields are often just as thin as letters or thinner.

I think we have to balance that with the reality of testing and the burden on authors. And also the reality that this applies to focus rings which gives us a third layer of contrast. The button has to contrast with the focus ring which in turn has to contrast with the background. Is there any combination that can do that.

Sure: white, black, white. I say that simply to point out that many statements are greatly exaggerating the "limitation" this criterion would put on authors. I'd much rather we take the time to define the "visual identifiers" of importance further, offer more examples, compute the palette limitations, etc. before making snap changes to the SC and "sky is falling" statements like:

Will marketing departments accept such a limited palette?

[DavidMacDonald]

I'd much rather we take the time to define the "visual identifiers" of importance further, offer more examples, compute the palette limitations, etc. before making snap changes to the SC and "sky is falling" statements

I'd say it the other way around. The onus is on us to prove its easy to meet this kind of requirement. I say it is not easy to get three layers of contrast, unless the focus is changing colors, or is a multi color focus ring as I'm trying to get into CSS . We have a precedence in WCAG of 3 layers of contrast. G183 is a technique to pass the use of color SC by having the link text have a 3:1 contrast with surrounding static text and a 4.5 contrast with the background. The example for G183 shows a limited palette even at 3:1.

I spend much of my time negotiating with marketing departments for large companies, and I think its not at all a "sky is falling" statement. They can spend $100,000 on their colour palette. This stuff matters to them.

@jnurthen what are your thoughts? It's your comment.

[alastc]

I’m not sure the testing aspect is understood? (in the same way at least!)

If a button has a background, that needs to be 3:1 against it’s souroundings, or it can have a thin outline that can contrast with either the outer background or inner (button) background.

For a focus highlight, I think that can contrast with the inner or outer background, or the border if there is one. It doesn’t have to contrast with everything. It also doesn’t have to be complete (all the way around the control), I think Glenda has some examples

Is that how other people have understood it?

[Ryladog]

Yes!

​​​​​ katie

[DavidMacDonald]

I have understood that the focus ring needs to be seen. Let's say its black and it is focusing on a control that has a black border. It will have sufficient contrast with the outside background, but it is not seen very easily because it is the same colour as the border. From what I understand, you are saying that it passes, because it is 4.5:1 with the outer background.

That seems less than useful.

[steverep]

I say it is not easy to get three layers of contrast, unless the focus is changing colors, or is a multi color focus ring as I'm trying to get into CSS .

This is where I think we need to start when talking about changing the SC... In nearly all cases, 3-way contrast is not necessary as @alastc has pointed out. For focus indication in particular, there are infinite possibilities beyond just the typical ring, and even when a ring is used it's not a 3-way requirement. And in cases where an author has backed themselves into a corner, it's usually incredibly simple to get out of it with a simple border or other technique.

I spend much of my time negotiating with marketing departments for large companies, and I think its not at all a "sky is falling" statement. They can spend $100,000 on their colour palette. This stuff matters to them.

My apologies - I did not mean your point was not important and certainly don't discount your experience. It's just evident that folks still don't truly understand the actual requirements, and that should start with better language rather than reduced accessibility.

[DavidMacDonald]

@steverep

In nearly all cases, 3-way contrast is not necessary as @alastc has pointed out.

See the above comment https://github.com/w3c/wcag21/issues/490#issuecomment-3372443390

border in CSS changes the layout, so as the user tabs around the layout jumps up and down a couple of pixels. The author would have to use outline. I'm trying to get more flexibility into CSS outline. I'm not sure the status of the that. It's been accepted in principle to allow more than one outline colours by core authors.

[steverep]

I have understood that the focus ring needs to be seen. Let's say its black and it is focusing on a control that has a black border. It will have sufficient contrast with the outside background, but it is not seen very easily because it is the same colour as the border. From what I understand, you are saying that it passes, because it is 4.5:1 with the outer background. That seems less than useful.

So, this is a simple scenario that perhaps we should take back to the LVTF quickly. I agree that it's not ideal, especially if the border has zero or very little contrast with the button background. The special think about what is happening here is that if the button starts with no border and then a border is added when focused, the content change also needs to be taken into account. How much of a content change is required to successfully indicate focus or selection for low vision is a much more difficult question given the full range of CSS possibilities, and goes way beyond this SC.

[joshueoconnor]

Survey https://www.w3.org/2002/09/wbs/35422/Oct_17th_agwg/results

[awkawk]

Done