Slashdot Mirror
Accessibility In Linux Is Good (But Could Be Much Better)
An anonymous reader sends this report from opensource.com: GNU/Linux distributions provide great advantages over proprietary alternatives for people with disabilities. All the accessibility tools included in Linux are open source, meaning their code is readily available if you want to examine or improve it, and cost nothing. Hardware devices, of course, are still going to cost money. Additionally, accessibility software on other platforms generally contain licensing constraints on the user. ... When it comes to accessibility, Linux is not without issues. ... The number of developers who specifically work on accessibility tools is quite small. For example, there is only one Orca developer, two AT-SPI developers, and a single GTK developer. ... Developers who do not depend on assistive technologies tend to forget—or don't know—that a disabled person might want to use their application, read their web page, and so on. ... The problem is not necessarily that developers do not care. Rather, it's is that accessibility is highly specialized and requires someone with knowledge in the area, regardless of platform.
...there is only one Orca developer, two AT-SPI developers, and a single GTK developer.
Really? Only one person develops GTK? Who'd'a thunk?
why not?
for the sake of children and disabled
The reality is, most developers have a fixed and small amount of time to develop product. They aren't going to work on features for 2% of people when features that 50% of people can use need work.
Mean what you say...say what you mean.
Just because it's open source doesn't mean it's great. I'd classify accessibility for blind/less sighted users to be non-existent (with the exception of a few applications). Every iteration of X Windows since X11 has been worse and worse with its implementation, and if things go wrong it is nearly impossible to get around. A few applications that implement the full GTK stack /properly/ are passable, but those that use Gnome's or KDE's tools don't pass text back to a speech engine at all.
Console is fine, but as soon as you try and use a tool that uses ncurses or any other menuing application you are SOL.
Firefox hasn't worked well with a screen reader in about 5 years. Never was able to get Chrome fully installed.
However, the adoption of Linux within workplaces can certainly be constrained by, for example, ADA requirements. The lack of proper accessibility may ultimately prevent certain businesses or organizations from implementing Linux when it would otherwise be most preferable, simply because it does not satisfy their need for compliance. I'm sure you can see the potential ripple effects from that kind of restriction and how it might impact even those developers who do take accessibility needs into consideration.
I read both his articles, and they're actually about "which development environment for disabled users is more open-source friendly." What the fuck? Obviously, Linux is going to be, so what? And there isn't a single word about which is actually better for disabled people to use, so provide great advantages over proprietary alternatives for people with disabilities is an inaccurate way to summarize the article. Instead, it's some guy blindly pursuing some nerdy "open source is the best!" dialogue like it was 2003.
Slashdot: providing anti-social weirdos a soapbox, since 1997.
How accessible is the best open source dev suite?
"All the accessibility tools included in Linux are open source, meaning their code is readily available if you want to examine or improve it"
This "it's better because you can fix it yourself" is usually pretty dubious. In this case, it's worthless unless the folks who need the accessibility can work on it. What's the most accessible dev environment? Are its accessibility features usable? Does it support all developing all the tools that need improvement?
If you have to get one, try and make it a common one.
that the Republicans hate us and want us to die. They ignore us then shit on us. That is their way. They hate us. That is why their Internet is so hateful to us. Just commenting here is a pain since this hateful site requires us to download a sound file and play it. Why do the Republicans that rule this site do that unless they hate us?
And the reality is, if you want contract with the government (such as universities), you need to make content accessible when relevant to the job. If not, the government will impose fines, and at best, just not use or renew your platform.
Money is a great motivator.
Unsponsored open source developers do so mostly because the projects they work on scratch some itch they have. If none of them are itching for accessibility, then there's no scratching.
GNU/Linux distributions provide great advantages over proprietary alternatives for people with disabilities. All the accessibility tools included in Linux are open source, meaning their code is readily available if you want to examine or improve it, and cost nothing.
Because disabled people are so looking for a DIY solution. I'll give you the one about cost, but aids for the disabled are usually sold or given away far under their actual cost due to ideal organizations, corporate PR, government aid and so on unless you're making a business specifically for the disabled.
Developers who do not depend on assistive technologies tend to forget - or don't know - that a disabled person might want to use their application, read their web page, and so on. ... The problem is not necessarily that developers do not care.
Oh please, the open source community is 95% driven by scratching your own itch. Very few do any real effort to make it easier for other people to use in general, disabled or not. Which of course doesn't mean that we're heartless bastards, we do care that there are children starving in Africa and a blind guy who can't use your app. Just not enough to ever get around to it.
Rather, it's is that accessibility is highly specialized and requires someone with knowledge in the area, regardless of platform.
Yes, but it's usually not rocket surgery if you care enough to explore it. The few times I've dabbled in it I've found that often takes a lot of effort that doesn't benefit anyone but the disabled, the way a wheelchair user needs a ramp where a step works fine for everybody else. Or in other words, even if you know what you're doing it still takes time, that I certainly wouldn't want to spend on a hobby project.
Live today, because you never know what tomorrow brings
Unlike proprietary alternatives...Linux distros with the Gnome desktop...includes accessibility tools out of the box, such as:
Screen reader A text-to-speech system to read what's on the screen
Magnifier Helps users with visual impairments who need larger text and images
High-contrast mode Helps users who have trouble seeing text unless contrast is corrected, such as white text on a black background, or vice versa
Mouse keys Controls the mouse using the number pad
Sticky keys Helps users who have trouble pressing multiple keys at once, and users who have use of only one hand
Bounce keys To ignore rapidly pressed keys or if a key is accidentally held down
On screen keyboard Helps users who cannot type at all, but who can use a mouse Visual alerts Replace system sounds with visual cues
Accessibility in Linux is good (but could be much better)
Compare:
While this article is aimed at Windows 95 much of the information on Accessibility Options also applies to Windows 3.x and Windows 98.
Making Windows 95 Accessible
It is much easier for a blind person to navigate a GUI if all of the widgets have meaningful names. I also makes it much easier to design test automation for that SW. Hint: it is not easy to navigate a stack of windows all of which are named "Window".
"code is readily available if you want to examine or improve it"
Means nothing if you aren't a developer at that level. It's just impossible for any random person who is not an experienced developer to jump in and modify code successfully.
I wish they'd stop saying this. Anybody to whom it applies already knows this, and to everybody else it inspires the false hope that any random FOSS project they encounter will one day be usable in a way they'd like it to be. False.
"Select the following link to learn more about the Philadelphia Computer Users’ Group for the Blind and Visually Impaired" ref
Could someone explain to me why Linux users are in complete and total fucking denial about the reality of their favorite OS/Desktop/Distribution/Whatever? Fucking christ almighty. As has been pointed out, Windows has all these features and they work.
Things like accessibility have been hit or miss in distros for a while now. Although, I think it has more to do with the overall design issues rather than the number of people working on the projects.
Someone already commented on the GTK vs. Gnome vs. KDE issues. The fact that the window manager is responsible for a vast portion of the accessibility features in linux is a bad design choice. There is absolutely ZERO assurance that a given system will have a given window manager installed, or that a program will have a specific API available beyond what is provided by the standard libc and the kernel's syscalls. This is by design in linux and is really the only assurance given to a programmer. As such, a would be application developer has a choice to make, either abstract off the calls to the accessibility layer and code for each possible layer in existence to ensure compatibility OR hope that the window manager is smart enough to figure out that stuff for you, that the window manager's requirements for that functionality are installed and working, and be completely dependent on whatever window manager's API is chosen for use by the program. Both of these fail for the following reasons: The first is too difficult and time consuming to implement by a single project, and even if such an attempt were to succeed, it would only continue to be successful as long as any changes to the upstream accessibility layers could be kept up with. The sheer number of layers to support makes this an almost impossible task to maintain long term with out a huge community effort. The second option fails because of the original issue. THERE IS NO STANDARD API TO USE, AND NO ASSURANCES FOR WHAT IS AVAILABLE ON A TARGET SYSTEM. Just because KDE is installed (or whatever else you use) does not mean that KDE's accessibility code is fully working and that the API calls for whatever version of it the developer chose to use are present. (We have at least 2 major versions of KDE and GTK to support. KDE3 and KDE4. In addition to GTK2 and GTK3.)
For a real life example look at the state of input method editors on linux. Two major competing IME APIs. (IBus and SCIM) In both instances they use plugins to support the various major window manager APIs. For each unique window manager API in use on the system, a plugin must be installed for that specific window manager's API, or any program that uses that specific API will NOT have IME support. Which from the user's perspective, means that some programs may have IME support, others may not, and the user experience is broken. A user CANNOT expect that if the IME works in one program it will work in all of them. This is before you get into specifics however, such as IBus' lack of console mode support or the fact that IBus cannot support different window manager API versions concurrently. (GTK2 vs GTK3 is noticeable. Firefox is GTK2, but Libreoffice is GTK3. You can't run both with IBus support without some heavy scripting and environment hacking. Which is what Fedora does to make it all "just work", while distos like Mageia expect you to screw around with bash scripts and only use one program or the other at a given time.)
In reality it would be better for a standard API to be implemented across the board (that includes the console too not just the window managers) that can act as a sink for that information and pass it to the programs that can process it. (Whatever they may be.) The unfortunate thing is getting that done without forcing it into the kernel (or *shudders* systemd) is unlikely to happen. Standards are not easy to implement in linux by design, in keeping with it's tinkerers mentality. Although I do wonder why they can't agree to make everyone's lives easier for their own sake. (One API is easier to code for than 6 or 7 different ones, and leads to less bugs to deal with.)