Our great sponsors
-
-
WorkOS
The modern identity platform for B2B SaaS. The APIs are flexible and easy-to-use, supporting authentication, user identity, and complex enterprise features like SSO and SCIM provisioning.
-
-
> This entire thread started off by someone saying untrue rumors about Wayland and DE's regarding accessibility
I'm going to tentatively push back on this, my (very limited) understanding of the at-spi protocol is that it exposes application-specific controls and that it's not relevant to conversations about how to control a DE, and that current DEs do not in fact expose a common shared protocol across all of them for the kinds of controls that Talon needs.
I'm not an expert on this stuff by any means, but from what I can tell, OP is correct and is not spouting untrue rumors. I'd welcome correction in the form of some kind of accessible documentation for integrating with those window managers that uses a universal common protocol for all of them, but getting back to something I've pointed out multiple times above:
> so excuse everyone who thinks the complaints are likely misguided. [...] This is not rocket science...
This kind of response isn't particularly convincing given that I note you still haven't linked over some novice-accessible documentation about how to work with this stuff or test it, likely because that novice-accessible documentation is difficult to find if it exists at all.
I would love some easy ways to verify what at-spi controls, how to integrate with it in different languages, and how different DEs handle exposing it for their own interfaces. The absence of that accessibility is a big reason why people are confused about Wayland and might go a long way to explaining why application developers aren't quickly modernizing.
If you think that Talon developers are lying about what Wayland supports, then education efforts and clear, user-and-developer facing documentation about the protocols involved would be a pretty good way of combating FUD.
----
> Right now XWayland works as it should. So provide some specifics or get off the pot...
First, as others have mentioned, XWayland is insufficient for the kinds of needs that Talon has; operations like enumerating windows or handling focus can not (and should not) be handled via XWayland unless you plan to run your entire desktop environment in XWayland.
Secondly, XWayland is not the future. It is also a dead end. I'm glad that it exists for now, it is absolutely critical for early adoption of Wayland and is critical for supporting some apps. But XWayland is not a permanent solution to accessibility problems and it's wild to hear people simultaneously say that X11 is a walking zombie and to tell people that accessibility is fine because if there are any problems with Wayland you can always just tie your applications to a walking zombie.
The goal here is to get rid X11, or at least that's what I think the goal should be, because I do think that X11 is a walking zombie and I resent the fact that I'm still using it.
----
My concern is not that Wayland can't be accessible or that there aren't ways to make it accessible. My concern is that regardless of whether or not universal protocols exist right now, it doesn't look like they're being used and it doesn't look like Wayland is going to be accessible for a lot of people in reality. And I don't care who's fault that is; it's still the case that calling developers lazy and accusing people who are sharing real accessibility problems that they have right now of spreading "untrue rumors" does not seem to be helping to improve the situation.
This isn't a situation like NVIDIA where there's one specific company we can all shame; if a bunch of disparate developers aren't taking the necessary steps to update their applications, then something is going wrong in the communication process or the steps aren't clear enough, or the incentives aren't aligned.
Again, I would welcome clear and accessible documentation that average developers could use that demonstrates how Talon could be made to work with Wayland and how the concerns its developers have (https://github.com/splondike/wayland-accessibility-notes/blo...) could be handled.
Talon is what I would describe as accessibility for programmers (my definition, not the developer's!). You effectively write software that replaces keyboard and mouse usage, generically, flexibly and programmatically.
So when you write commands you bind them in different way: app specific[1], feature specific[2][3], OS-specific, hecking __programming language specific__[4] etc, and the in talon mixes and matches all of that stuff together.
So let's say I have VSCode focused on a javascript file . Talon knows this, and so I have "panel switch" which is a vscode specific command, and "op strict equal" to insert ` === `, but I also have generic text editing commands (because it's an editor), and multi cursor commands (because vscode has been tagged as multi cursor supporting), and tab commands (because vscode is a tab-based editor), and so on and so on.
If I then switched to the browser I would keep the generic text editing commands, and the tab commands, as it supports both of those things, but I would no longer have multi cursor support (or JS commands), because my browser doesn't support that.
This also means you can by and large use the same talon config (and so the same voice commands) on windows, mac and x11.
So for me switching to windows is actually less of a pain because most of the ways I interact with my computer don't actually change, as talon abstracts that away quite a bit.
[1] https://github.com/talonhub/community/blob/main/apps/vscode/... / https://github.com/talonhub/community/blob/main/apps/vscode/...
I want to move to the "future", but I use several firefox profiles via xpra in combination with xdotool based hotkeys. The fluidity with which I can control my different profile windows without a mouse and with which I can switch between computers with no lag with this setup is a big reason why I have stuck with a Linux desktop, so I am sad to see those super powers go away for no perceived benefit. I have tried Wayland several times now and don't notice any performance differences, just more bugs in Wayland.
`xdotool search` seems like it has been deemed a security issue for reasons I can't understand (if someone has hacked in to the point that they can even run such query commands, surely I'm already pwn'd). Maybe I'm oversimplifying it, but to me it's like being upset that someone who broke into your house can see the color of your curtains. The powerful feature set of X far outweighs these minor security concerns.
And it looks like xpra is facing huge issues switching over https://github.com/Xpra-org/xpra/issues/387 :(, I have yet to be as satisfied with any other free remote desktop software (paid nomachine is close but less scriptable).
I hope these aren't fundamental limitations of wayland, but the challenges seem steep.
> This is very true, and unfortunately there are very few people working on linux accessibility (including not me! I am part of the problem!).
Accessibility work itself ironically suffers from an accessibility problem. I brought up i3wm above, the issue for that is pretty illuminating: https://github.com/i3/i3/issues/3393
It's not that the devs are saying "this doesn't matter", the devs behind one of the most popular tiling window managers in the X11 ecosystem are saying, "this does matter, but we don't know how to fix it. We don't know what changes we'd need to make to get Orca working."
It's a really fundamental breakdown that's kind of a tragedy because I honestly believe that if accessibility communities were more heavily baked into testing and development in Linux and if this wasn't treated like two separate worlds, it would be better for everyone -- fixing accessibility concerns very often improves interfaces across the board and makes them more powerful.
But... how do you bridge that gap? I don't really know, I tried looking into Orca to see what would need to happen here and bounced off of it pretty hard, it's not a very approachable tech stack and there aren't tutorials or getting started guides. And on the other side of the issue I can preach about needing accessibility input during interface design, but I'm not in a position to give specific advice because I don't use screenreaders or alternate control schemes and I don't know what the biggest problems are.
The people who need to be involved in that process can't get involved because there's a tech barrier in place even for technically inclined people, and because the underlying software locks them out from the start. i3wm isn't ever going to get someone who's intimately familiar with Orca to jump into the conversation because the people who need to use Orca can't use i3wm. So that leaves the people who can address that tech barrier, but they don't know what to do or how to approach the problem because of the lack of involvement and because the communities are isolated from each other. So it's a chicken-and-egg problem and I don't know how to solve it.