RE: on-page controls considered bad practice?

This has been one of the accessibility community's "holy wars" for two decades. My view is that designers are perfectly aware that these features exist in browsers, but they are buried so deep (and in different places in different browsers) that users do not know they exist.

I interview lots of people with disabilities prior to their participation in user testing, and almost none use the browser's built-in accessibility features even when they would clearly be beneficial. One of the issues is that they don't necessarily take effect on all websites, depending on the coding. In some cases, they break perfectly good websites.

I am therefore an advocate of having on-page controls, with the caveat that they must be easy to use. I have seen many implementations that were overly complex. That includes some of the third-party accessibility overlays.

There is nothing in any accessibility standard that prohibits the provision of on-page controls. I would say you should not discourage their use unless you have evidence to support that view. As Deming rather harshly said, "Without data, you're just another person with an opinion".

Steve Green
Managing Director
Test Partners Ltd


-----Original Message-----
From: Tobias Bengfort <tobias.bengfort@posteo.de> 
Sent: 09 March 2021 13:41
To: WAI IG <w3c-wai-ig@w3.org>
Subject: on-page controls considered bad practice?

Hi,

I often see websites that have on-page controls for things that can already be control via browser settings, e.g. font-face, font-size, light/dark theme etc. I recently even discovered an on-page screen reader.

While these controls are not harmful by themselves I consider them red flags because they could imply that the designers were not aware of the native controls. I usually give the feedback that I consider these controls unnecessary. But I have never seen any authoritative guidelines on this topic.

Do you know of any regulations or articles on that topic? Am I right in discouraging these controls or are there valid usecases?

thanks
tobias

Received on Tuesday, 9 March 2021 15:08:26 UTC