Use of screen reader or assistive technology detectable?
With iOS 12.2, Apple had introduced a feature that notified websites and apps that a user was using assistive technology, especially the screen reader VoiceOver. This feature is unspectacular in itself, had it not been enabled by default setting and hidden deep within the system. Apple dropped this feature, which was allegedly never properly enabled, after numerous protests. This article is about why the detection of the use of assistive technology for website operators is rejected by many disabled persons.
The Electronic Frontier Foundation showed some time ago that in principle every system can be identified by just a handful of data. Almost every system has a digital fingerprint.
- Browser side detection
- The advantages
- The disadvantages
- Conclusion - my screen reader is none of your business
- Further reading
Browser side detection
However, browsers usually also know that the user is using assistive software, at least this is true for screen readers. Since screen readers rely on certain information from the accessibility API's, the browser must convey this information. However, I don't know what the situation is with magnification software or eye tracking and similar systems. These programs require much less information than screen readers. However, the use of assistive technology is not passed to the server. The browser retrieves all information indiscriminately, whether a screen reader is installed or not.
There are some advantages. For example, the tiresome discussion about whether and how many disabled persons use a website would be eliminated. If the browser passes on the info to the server statistics, we know it, can put it into numbers and beat it around the ears of our unwilling CEO, who thinks accessibility is expensive devil stuff.
An obvious advantage is that a server could deliver a more accessible version of the website if it knows that the user has a disability. It is conceivable, for example, that the server could receive the information that the visually impaired person needs a color combination of white on black and a minimum text size of 20 pt and deliver an appropriate version. The deaf or learning disabled could receive a version in sign language or easy language and so on. However, some of this stuff is already possible with media queries, but almost nowone uses the accessibility-related media queries.
However, these ideas are rejected for good reasons. The consensus today is that the best solution is to make the website itself as accessible as possible and leave everything else up to the user, their assistive software, and the browser properties. The best website in the world could not cover all possible visual impairments with special versions. The easy-language version - which, by the way, is mostly as missing as content in sign language - can easily be made findable at the top right of the website, as these groups of persons have learned to do.
I don't see any advantage for other groups either. Autistic persons or persons with epilepsy can set their browsers to minimize interference from animations, and they usually don't use more advanced assistive software anyway unless there are other disabilities.
This also defeats the statistics argument. We may know at the end of the day how many screen reader users visited the site, but most disabled persons don't use assistive software at all, they just change a few settings.
And even if the numbers reflected reality, little would change. We've been complaining about Google, Xing, or LinkedIn's lack of accessibility for years, so they should have caught on by now, but don't see the need for action. And, how many disabled persons do I even need before I consider caring about accessibility? Couldn't it be a counter argument when the descision maker sees that only 0.001 per cent of the visitors uses screen readers or magnifiers?
Another factor that might go under the radar: The whole thing could also be privacy-related. The electronic fingerprint mentioned above would be made a little more precise by adding another piece of information - auxiliary software. Let's say I send a complaint form via the provider's website, who stops the server from attaching the info about the assistive software so that the recipient knows he has a cripple in front of him? Am I being paranoid? Yes maybe, but it doesn't matter. Past experience shows that data parsimony is always better than verbiage.
Conclusion - my screen reader is none of your business
I think the point is clear. The info that I use a screen reader is really none of the web operator's business, it's irrelevant to them, and it doesn't give us any direct benefits that we want at this price.