The point of the article was not to pick on Twitter. While using a tool like AccessibleTwitter is great for that particular example, it's not fixing the real problem - people will run into problems with modern websites, and simple things can make that experience better.
I hope I didn't make a mistake by pointing at Twitter.
There's a Firefox extension called Fangs that I use, but it's not that great. I plan to have some posts soon that talk about tools.
JAWS for Windows gives you a 45 minute trial. All you need to do to restart the trial is to reboot your machine.
But like I said in the article - it's not gonna be a checklist that gets the job done, it's gonna be about testing it in real screenreaders. I haven't seen any emulators that behave like the real ones.
HTML / browsers are entirely capable of handling websites easily for blind people / other needs. The problem is that nobody builds their pages to use the tools already in existence - even really high profile pages are generally atrocious.
Similarly, take a peek around the web (or at most applications) and consider if you couldn't tell the difference between red/green or yellow/blue. Most people just don't consider it if it isn't shown to be a problem, and yes, it does take more time to develop for those edge cases.
I was part of a small team contracted to build the screenreader for Vista (Narrator). It's a tough thing to get right. Once we had a working version I tried finding and starting Visual Studio and building a simple Winforms app all with the monitor off. I didn't get very far.
I imagine you wouldn't get very far drawing a picture in MS Paint, either. Some tasks are inherently visual (like laying out controls in a 2D plane.) Do you think you could have managed to build a CLI app?
I once worked for a company who had to maintain 508 compliance (accessible to people with disabilities) for their flagship product as many of the clients were government agencies or agencies with government grants. Unfortunately this is exactly what our product was like with a screen reader, probably worse. Basically everything just needed to be accessible correctly, couldn't use a clickable div/span for whatever crazy reason just anchor tags. It's sad that the blind have to live with crappy software, I always felt bad for them when hearing from our managers to check for compliance a few times a week.
Maybe I'd get used to it, but the computer voice really bugs me and feels hard to understand, making it hard to even get enough sense out of it to get to the point of being confused by the content. Is this what typical screenreaders sound like? To my ears, the vocal synthesis in espeak is considerably more pleasant (http://espeak.sourceforge.net/).
Pardon me if this is an arrogantly undisabled thing to say but.. perhaps rather than making sites better, we could be making better screenreaders? Interpreting the truly important information on a page programatically isn't the easiest task but it's hardly a freakishly crazy AI problem either.. Any startups in this field? :)
I don't know how I can answer this. I suppose I could make a comment about how we don't even have browsers for the sighted that can understand the code that some developers throw up on pages.
Or I could say "Find 10 pages, and quickly determine, from only the HTML markup, what the main content is." and then do that for the top 1000 pages on the net. See how hard the problem is. If you solve it, you'd be a hero.
On the other hand, you could use accessibility patterns like UJS, alternative text, skip links, ARIA roles, better semantic markup, and make your site about 80% more usable in about an hour :)
On the other hand, you could use accessibility patterns like UJS, alternative text, skip links, ARIA roles, better semantic markup, and make your site about 80% more usable in about an hour :)
I agree, but sadly it remains easier to solve even hard technical challenges than seemingly easy social ones where, if you're lucky, perhaps 20% of developers will cooperate.
As a developer I find accessibility interesting, which is why I submitted it, but unless I'm being forced to implement it by law or context (e.g. a site with a high demographic of people using screen readers) I haven't the inclination or motivation to focus on it (though skip links, alternative text, etc, are good ideas anyway). It's a bit like dealing with IE5 or WebTV in that regard.
There are lots of ways in which we could make the world better for disabled people, I'm sure, but typically we've built technology to make life easier for them instead of changing the way the world already operates (with the notable exception of building codes enforcing the installation of ramps and disabled access systems). I'm not saying that's always a good thing but people are notoriously slow to change en masse (y'seen the average quality of HTML out there?).
It shouldn't be too far fetched to significantly improve the usability of screen readers even on the hairiest of pages - most regular browsers do a stellar job at interpreting horrendous markup. Half of the stuff that your recording comes out with isn't even visible on the screen.. do screen readers pay full attention to CSS and states enforced and controlled by JavaScript? If not, that's probably step one to making things better.
I suspect the problem is that there's little money, fame, and too few people looking at the technological aspects of this problem, whereas inventing a new "standard" and sitting on some W3 fancy board looks great on the old résumé.
21 comments
[ 5.2 ms ] story [ 53.6 ms ] threadCan I ask what OS you're working with?
The full video is at http://napcsweb.com/files/video/01_screenreader/01_screenrea... or http://napcsweb.com/files/video/01_screenreader/01_screenrea... if you want to watch it.
(edit: It worked for me even before the vimeo switch)
I hope I didn't make a mistake by pointing at Twitter.
That'd make it infinitely easier for web developers to test pages...
JAWS for Windows gives you a 45 minute trial. All you need to do to restart the trial is to reboot your machine.
But like I said in the article - it's not gonna be a checklist that gets the job done, it's gonna be about testing it in real screenreaders. I haven't seen any emulators that behave like the real ones.
Edit: http://www.webtospeech.com is free btw.
HTML / browsers are entirely capable of handling websites easily for blind people / other needs. The problem is that nobody builds their pages to use the tools already in existence - even really high profile pages are generally atrocious.
Similarly, take a peek around the web (or at most applications) and consider if you couldn't tell the difference between red/green or yellow/blue. Most people just don't consider it if it isn't shown to be a problem, and yes, it does take more time to develop for those edge cases.
Or I could say "Find 10 pages, and quickly determine, from only the HTML markup, what the main content is." and then do that for the top 1000 pages on the net. See how hard the problem is. If you solve it, you'd be a hero.
On the other hand, you could use accessibility patterns like UJS, alternative text, skip links, ARIA roles, better semantic markup, and make your site about 80% more usable in about an hour :)
I agree, but sadly it remains easier to solve even hard technical challenges than seemingly easy social ones where, if you're lucky, perhaps 20% of developers will cooperate.
As a developer I find accessibility interesting, which is why I submitted it, but unless I'm being forced to implement it by law or context (e.g. a site with a high demographic of people using screen readers) I haven't the inclination or motivation to focus on it (though skip links, alternative text, etc, are good ideas anyway). It's a bit like dealing with IE5 or WebTV in that regard.
There are lots of ways in which we could make the world better for disabled people, I'm sure, but typically we've built technology to make life easier for them instead of changing the way the world already operates (with the notable exception of building codes enforcing the installation of ramps and disabled access systems). I'm not saying that's always a good thing but people are notoriously slow to change en masse (y'seen the average quality of HTML out there?).
It shouldn't be too far fetched to significantly improve the usability of screen readers even on the hairiest of pages - most regular browsers do a stellar job at interpreting horrendous markup. Half of the stuff that your recording comes out with isn't even visible on the screen.. do screen readers pay full attention to CSS and states enforced and controlled by JavaScript? If not, that's probably step one to making things better.
I suspect the problem is that there's little money, fame, and too few people looking at the technological aspects of this problem, whereas inventing a new "standard" and sitting on some W3 fancy board looks great on the old résumé.