Note: the following are my personal opinions, and should not necessarily be taken as the opinion of my employer.Back in the day, when I was a larval teenaged geek, one of the first projects I ever created was called mircspeech. It was a script for the then popular mIRC client, that interfaced it with the Jaws For Windows screen reader. This allowed any blind or visually impaired person to join any of the chatrooms of the day. The IRC protocol was used by the chat features of Microsoft, ICQ, and everyone in between.

Even as the 2000’s turned into the 2010’s, and the popularity of IRC began to wane, most popular services like slack and Twitch continued to support connecting with an IRC client. But that wouldn’t last forever. Discord never built an official gateway, Slack eventually shut there’s down, and using a standard IRC client with Twitch is a good deal harder than it used to be.

For blind people, this meant that instead of choosing the software that worked best for us, we needed to use the official apps for each service. Some of these apps were accessible, some became so, and some are still on the journey. Regardless, none of the apps offered the degree of customization allowed by services using open protocols.

We saw the same pattern happen with Twitter. Originally, Twitter offered an open and freely available API. While the official client and website had accessibility challenges, a plethora of more accessible clients sprung up to meet the accessibility needs of users: chicken nugget, TWBlue, qwitter, jawter, Tweesecake, and many, many more. But when the Twitter API was restricted and shut down, all of this accessibility work was lost. Once again, people with disabilities had to use the official app, and could no longer customize the service to their needs.

Most recently, the story repeated itself once more with Reddit. Specialty apps developed to meet the needs of screen-reader users like Reddit For Blind, Luna, and Dystopia were heavily impacted by the changes Reddit made to its API, pushing everyone to use the official Reddit app.

But that’s not to say it’s been all bad news. The mass adoption of OpenID for logins has meant that for many services, everyone can choose the provider that works best and is most accessible for them: Google? Apple? GitHub? Thanks to OpenID, we all generally get to pick our favourite. Personally, I like the way Apple Presents two-factor codes to me, so use that whenever I can; I find Apple’s two factor the easiest for me to read with my screen reader and act on. But if your milage varies (because everyone’s needs are different), thanks to OpenID, you can probably pick something else.

You’ll also find thriving communities of blind people on Mastodon and throughout the Fediverse. Thanks to the open standards, open protocols, and open API’s that the fediverse is built on, we have dozens of accessible apps to pick from. Some are specialized for our needs, and some are accessible third-party apps.

And although text-based usenet is almost completely gone, email remains alive and well, with dozens of accessibility communities making their homes in places like GroupsIO and FreeLists.

While the discussion about just how decentralized and federated bluesky is continues, it’s openly documented nature means that third-party accessibility apps for it are also slowly beginning to emerge.

I write all this in support of a controversial hot take that’s been percolating in my brain for a while now: open Protocols and APIs are the most important feature for robust and lasting accessibility to exist. In fact, true accessibility might not be possible without them. An API on top of a proprietary service can be shut down or changed, as happened with Twitter. Open source code is all well and good, but if it doesn’t implement a robust API-first design, the results of that code won’t be accessible; this can be seen most clearly in the current state of accessibility on the Linux desktop. Web Standards and open formats are, of course, an important part of accessibility, but if browsers didn’t implement API’s like IAccessible2, assistive technology would be dead in the water. When the protocol and API are both documented, though, apps quickly spring up to meet the needs of all users. And if the protocol is also open, as with email or IRC, the accessibility granted by those apps can never be taken away. User Adoption, of course, is another matter entirely. While I still maintain a project to make a modern IRC client accessible, IRC’s failure to modernize as a protocol means effectively nobody uses it. That makes its ongoing accessibility much less meaningful than I would like. I’m not sure what the solution to adoption is; but I know that just like everywhere else social problems and technology intersect, it will be complicated, messy, and on-going. Perhaps the only lesson of this article is that accessibility is as much (or more) a social problem as a technical one.