Is there a coop or collective of disabled people in tech where you can go to hire accessibility experts? I don't want to continue to release software unless I know its accessible, but I can't know that unless I have my software audited by disabled tech experts that know the pitfalls of these things. I don't need free labour either. I am happy to pay for this service.

#a11y #accessbility #disablity #foss

Sorry for the extreme delay. As I mentioned earlier, this is my hobby project, so I can only dedicate a lot of time to it when work isn't demanding on weekdays, but more frequently on the weekend! The project is called smol-k8s-lab (pronounced "small kay eights lab" which is the trick I used to get it to be pronounced properly by the tts library in python). You can install it using pipx, so pipx install smol-k8s-lab, but it has external dependencies for using kubernetes, so you'll also need to install with apt or brew kind and k3d. I have added screen reading elements for the terminal UI, but they're all prerecorded MP3s because I found orca and espeak to not sound very good. The TTS library I used with python has an irish accent, which I think is understandable, but I'd love your feedback on. Because the program uses MP3s, you will need alsa and/or pulse libraries installed to actually play the audio. If you would like, you can check out the project on github here: github.com/small-hack/smol-k8s…

If you have questions, or you would like to chat on a voice call, I can help out if we can find a time to link up. I can do weekends in the mornings after 10AM CEST, so tomorrow could work if either of you would like to do a call. I guess it makes sense to use something were we can share screens, that way if you run into any hiccups, I can see what you can't, and live fix any issues so you can continue. I've tested it form the point of running the smol-k8s-lab command, and then gone through the accessibility options. I am very interested in what you think of the accessibility screen and options and if I should improve them. I allow setting terminal beeps as well as reading various elements including screen titles. What I don't know how to do is make orca read the text. I know how to launch orca, and I know how to use the MP3s I added or espeak, but I don't understand how to tell orca, from the command line, to read a line of text.

Sorry for the very long reply. I know they're a pain on screenreaders, but there's a lot of info here and I didn't want to leave anything out. If you're still interested, please let me know what app you'd like to meet on and I'll get you my info. If you'd like to do it on your own and just provide feedback that's also ok. Please don't hold back any negative feedback. If you think it's an awful experience, tell me, and tell me how you'd like it to be better. I am here to help, as long as you understand that it may take me a few weeks as I do this in my free time. If you'd like payment for your services, please just DM me your rate and how best to pay you. I can do paypal through my roommate, direct bank transfer if you have an IBAN account, and ko-fi. The reason paypal has to be through my roommate is that paypal canceled my account when I tried to change my name after transitioning my gender.

Oh, this is also the first project I am confident enough to introduce to the blind community, but if you don't mind working with me, I have others that I'd like to work with you on in the future. I'd love to develop a relationship where I can start working more closely with people in the community that passionate about accessibility. I promise to learn from anything you tell me and disseminate that info back to my other friends in the industry and my coworkers if they'll hear it 🙇

Yes, it's in a full TUI right now. There's also a CLI that can be used, but the purpose of the TUI was to make it easier to walk you through the process of setting everything up. Technically, if you're good with YAML, you could just use the CLI, but then you'd need to fill out the yaml using only comments in the yaml file.

Do you know of a CLI command to play some text, so something like: orca say "welcome to the TUI"? I know how to do this with espeak, but I'm not sure how to make orca read a TUI specifically. Alternatively, if you can tell me the command to make orca read everything in the terminal, that'd be awesome, or even just point me at a guide where I can learn myself. I can also spend sometime this evening reading this guide I found, but I'm not in a space where I can do that right at this exact moment. github.com/C-Loftus/orca-intro…

I really appreciate it, but please don't stress yourself too much on this. If you don't have the time or your system isn't cooperating, you can always wait till next weekend.

Until then, I'm going to keep fighting pyobjc on macOS. I think part of my issues here are because I wrapped tdsr in a poetry project so I could develop with it a bit more quickly and then released it via pip. When I do a pipx install tdsr2 it doesn't work and says it can't find the objc library, but when I use poetry shell to create a local python virtual environment and use the tdsr command, it works beautifully. When I try to run smol-k8s-lab in the terminal after tdsr is running, it both plays the MP3s I mentioned before AND tdsr uses the default system TTS voice and you get both playing at once which is horrible. It also causes tdsr to start spitting out a lot of terminal escape key garbage to the screen and breaks the whole session entirely. I had to manually close my terminal tab, because it wasn't responding to control + D like it normally does. If I wrote the thing and I feel frustrated, I don't want users to also go through that.

As far as linux is concerned, I don't have easy access to a machine console with gnome on it, as I'm currently running KDE and tdsr doesn't seem to play well on kde at all. Let me give speakup a try though. I haven't tried that one yet. It looks like it uses espeak, which I have tried before, so maybe it'll work better. 🙏 (The other thing that makes this complicated is supporting both Linux and MacOS at the same time, because there's different tools across bother of these. I wonder if this is made any better by anything the wezterm project has done lately as that's a terminal that works on multiple OS. I brought this up to wez himself in 2023 github.com/wezterm/wezterm/dis…

I followed up with wezterm and it looks like he opened an issue where he's gotten about as far as I have in learning the tools he needs to learn, but hasn't put in any actual PRs github.com/wezterm/wezterm/iss… Having a single terminal to support across all operating systems though would make this easier, so I will push on this as well in the background.

I should read it soon

I am a hobby systems programmer that wants to make accessibility better but I don't really know how to since nobody besides me and maybe a few others seem to actually want to make Linux better beyond their pristine abled usecase

so far it has been an amazing read! I had no idea about all the accessibility tools.

I am working on a Linux distro (slowly) and would love to at least try to make an effort to solve at least some of this... Esp since I will be maintaining significant patches to the kernel, and handrolling my own DE, package manager...

I was reading this and thinking "hm, I wonder if NixOS might be a good platform to pursue for accessibility". Then I got to the section you had on NixOS.

It definitely seems like a good starting point. Some sort of configuration as code approach, anyway.

Though I think it'll continue to be a power tool that's hard to use for a long time, maybe forever.

> You're not just using Linux anymore.
You’re wrangling demons.

Anyone sufficiently far into Linux is not using Linux, but rather wrangling demons

> Not with fragile scripts or one-person distros—but with reproducible systems that treat configuration like infrastructure.

I switched from NixOS to a one-being mess of wrangled demons with hand-rolled Nix-on-Arch stuff with Fedora packages and random stuff shoved into /usr/local and a bunch of handrolled containers and aaaa

And I am sighted, and the reason why my system is the way it is is probably due to my attention disorder and other various needs.

I hate Linux so much that I have spent hundreds of hours on OSDev to replace Linux with my own OS haha

I have a few rants about glibc (nss and locales; vanilla ikik), idk brltty, and I have countless about the kernel.

I do need to learn more about the existing accessibility tools tho.

Wayland is such a PITA, even for sighted people. So many basic things they refused (and sometimes still refuse) to even acknowledge - IDK how it got any traction.

But at least for now using X11 is still feasible (I'm using it with XFCE myself)

omg I would love help, I have been fighting LLVM for the past 3 years haha

I wanna avoid glibc due to issues with static linking and code quality and instead wanna fork musl

Same with libstdc++, GCC...

It's pretty fucked up how apparently accessibility was not even considered when developing/specifying both Wayland and Gtk4, though.

And those are not exactly hobby projects developed only by volunteers in their freetime either

urgh - again a case of >90% there, but no one does the last 10%... how hard can it be to at at least document the env vars?!
(I vaguely remember trying to figure out some Qt env vars a while ago and didn't find any proper documentation either, so that's a general problem there)

And wouldn't it be nice if toolkits and desktops could agree on *one* (hopefully simple) way to enable such features...

These problems are frustrating because without a screen, sighted users can't use a computer either. Hopefully braille and text-to-speech will be treated as first class citizens one day.

A question. How practical is a full text mode linux system for blind users, at the moment? I know edbrowse exists, and standard command-line can do most other things. Is the main problem running specialised programs that only provide a GUI?

this, is, awesome! goddamn, a very, very good article!

However, and consider this the well...but, section. Essentially, a lot of your issues happen because ubuntu exists, and particularly because you're using it, especially ubuntu mate. Ubuntu, to me, stopped mattering the moment canonical did their whole snap thing and slowed down their system, not sure if that's still a thing, but I know they're using that flutter installer which is...ahem, definitely not a very good thing for us. I didn't try it, but I know from trusted sources that it's pretty bad, I wouldn't have the patience for that tbh.

Next, the login issues...stop using ubuntu mate, and in particular, the issue here is a display manager called lightdm. It's possible to make it work again by editting configuration files, but meh, I don't care at all at this point, use gdm is my recommendation. I know it's bulky and a bit meh, but it's bad for everyone, as someone once said. With gdm, if you start your screenreader once, it's gonna be on pretty much for ever, unless some specific thing happens, but that thing is so rare that you shouldn't ever encounter it. Also, persistance, you don't need a script for that, that hasn't been a thing for me ever since the dark ages. I'm using arch, but like, if people's first experience of stuff is ubuntu and especially ubuntu mate, no wonder they run in terror at the whole brokenness of it all, let's just say ubuntu is even more known for not caring

About pipewire: the way to make that work is to use systemd's support for linger and enable that, then indeed as you said, you share sockets with root, but you don't have to modify service files or session files or anything of the sort. If I would have to do that, honestly I would have given up a long time ago. I like challenges, but challenge for challenge's sake isn't something I'm after. O also, for configs, use the user wide location, $XDG_CONFIG_DIR/pipewire.conf.d, that wouldn't be overridden by updates of pipewire. I have a script for that, if you need it. It's still using the old lua based configuration for wireplumber, but that's not needed anymore anyway. Also, a lot of gtk4 apps just work nowadays, way fewer accessibility issues than before.

About alsa, the kernel and espeakup, no comment, unless you do what you said and what I also said above, it's kinda oofed and screwed. About brltty and friends, absolutely no idea there either, I don't have a braille display.

You mention the thing that if one doesn't know the secret command to turn on the screenreader, they're screwed. True, but there the only winner is mac OS, because unless you know to press ctrl+windows+enter, you won't get speech in windows. Is it consistent and all distros have it? hell no, definitely not! but just pointing that bit out.

Alright, that being said, if you want to know how a good linux future would look like, can you do the following? no problem if you don't want to, but still

• go on the fedora website
• download the fedora workstation iso, important, workstation, v42
• boot into it
• this issue is unescapable for now unfortunately, but when you think it booted, press alt+super+s to do the screenreader command
• the voice would fluctuate weirdly in pitch, but either way, try to just install it as usual, without going in and editting config files, or doing anything a regular user wouldn't be able to do. If you want to do that, can you please reply with how that went? I'm really curious, because I got a relative newbie to tech to install it with this new installer without issues, and yes, this is specific to 42, and yes, I think only workstation has it

Yeah, unfortunately, customisations aren't so huge with fedora and gnome, but yeah, that's the quickest way to get started. There are also some gnome peculiarities, but I'm assuming you know them, so I won't repeat those.

About the whole wayland stuff, totally agreed! however, starting from the premis that any app can snoop on my keyboard all the time, can read every single file I have access to, and can perform anything I can, and maybe more if it can become root, that's definitely not a good situation either. I hate the whole accessibility is left by the wayside as much as you do, hell, I'm making a screenreader along with some awesome people specifically because we want to be free to just experiment and maybe write something better than orca one day, aka what commentary is to talkback in the android realm, and we can barely figure any of this stuff out, we didn't even get to implementing orca's heuristics because we don't know what they are, since you guessed it, they aren't documented.

And yeah, I can't wait for your next article, it's gonna be awesome to read it! thanks for the very insightful text once again, it was really well written!

I may reply to some other posts of this thread because there are some things to still be clarified, don't like how the fediverse represents threads in a linear fassion, but yeah, telling you this so it wouldn't be confusing afterwards, because I was indeed confused when I saw this issue for the first time.

At the desktop environment level, KDE eV is currently employing/contracting the legally blind Ritchie Frodomar, 1 of 11 eV staff in total, which should hopefully help improve this situation. Still, it's awful that many Linux distros have currently got themselves into this state.

Note to self: I must install a screen reader on my GNU/Linux system (1) to know what the experience is like, (2) to be able to report/fix the bugs, and (3) to be able to test the code projects I will create.

I recall someone here on the Fedis shared a sad thread years back, about the two (2) times that the biggest American charity for blind folks partnered with serious Linux devs to improve accessibility in the core OS.

Both times, Microsoft threatened to defund the Charity, who received most of their money from MS, and the project fell through.

MS wins gov contracts on accessibility grounds, so they sabotaged the open alternative twice.

holy shit that's depressing. at least NixOS lets you share the horrible crimes you have to do to get things working but this is little assurance.

if you want reviews on nixpkgs changes, feel free to tag me (LF- on GitHub) but i don't know how else to practically help; it feels like basically everything is undermaintained, nobody has any time, and that the reason stuff worked in 2012 was that there were government contracts to be had for gnome people at redhat that required that the accessibility works.

ultimately all those things just set environment.etc.* so you can just do that there as well with raw text; the modules that give you something nicer are just implemented on top of that. it's also possible to manually create it if you don't want to manage it with NixOS.

in principle it's probably possible to use another distro's ISO to install NixOS. it's *very* possible to install NixOS over ssh with a custom ISO and i do it most times. building custom ISOs with your ssh keys built in is like 20 lines of code on *any* distro with lix/nix installed and it's my preferred installation mechanism; see LF-/dotfiles on GitHub, the iso stuff in configs/nix/flake.nix.

the iso not being accessible really sucks and i am willing to lend social capital to efforts fixing it; the solution of adding ssh keys to the installer can bypass that for now, and probably is more pleasant than doing an arch style install process entirely in a vt console

I saw your blog post passing by in another toot, and wrote a mini-thread of three toots about it: exquisite.social/@labellaragas…

I work in website accessiiblity and I am very sorry desktop accessibility on Linux is failing people with a disability. Are there distro's that do better as Ubuntu, or are they all bad?

And have you tried any of the BSD's?