Accessibility 2.0

Julie Howell will be moderating this panel discussion. Before that she has a few words to say. She sees Web 2.0 as an opportunity. Everyone is saying that social network sites are going to get more “vertical” and be based around niche interests — well, the accessibility community has been dealing with niche interest groups for years.

Here’s a big news announcement: PAS 78 is being turned into a British Standard. Consultation on the draft will take place around September.

Julie once again reiterates that Web content guidelines are becoming less important and authoring tool guidelines are more relevant.

Most of all, accessibility is not an old-school attitude. We should be providing rich user experiences to everyone.

Now to introduce the panelists: Mike Davies, Kath Moonan, Bim Egan, Jonathan Hassell, Antonia Hyde and Panayiotis Zaphiris.

Julie starts by asking Bim if Web accesibility is getting better or worse. Better! Bim is quite adamant. If you take a step backward, you’ll see that we’ve come on in leaps and beyonds. In fact, sometimes it goes too far: square wheel building. That’s when developers get over-zealous about accessibility and add “features” that do more harm than good.

Julie asks Mike why someone from Yahoo is at this conference. Mike says that it’s because they listen to their users although the reason why he’s here is probably because of his work at Legal and General where he demonstrated the business benefits of accessibility. Yahoo put together a kick-ass team, thanks to Murray. That’s why Mike is at Yahoo. Mike says the accessibility success stories are down to one person with power taking a stand rather than evangelists in the trenches.

Julie asks Antonia how we get people to think more about learning disabilities. Antonia says it’s about getting everyone together to collaborate and learn.

Staying on the subject of learning disabilities, Julie asks Jonathan about the reluctance of people to participate for fear of being seen as different. That can be a real problem online with its anonymous culture of flaming. The viewpoint of disabled people needs to be represented more.

Julie talks about Second Life. On the internet, we can theoretically transcend disabilities from meatspace. Jonathan says that escapism has its place but he wouldn’t want anybody to lose their appreciation of their identity. In a room of deaf people, not being able to communicate in BSL is a disability.

Julie wants to know what Kath thinks of PAS 78. Before getting on to that though, Kath would like to slam Yahoo for having inaccessible gateways into various services. Sorry Mike, sorry Christian, but Kath is handing you a plate with your ass on it. Kath then goes into a long rant and ends by pointing out that when she is testing accessibility she finds that what she’s really doing is testing usability. So to answer the question, it’s a lovely document. Seriously though, she says we need to be more agile with accessibility and PAS78 looks like it will be a standard before WCAG 2.0 is out. That’s a good pace.

Panayiotis is asked about the future and how the elderly population will cross over with disabled users. Before answering that question, he’d like to take up Kath’s point about being more agile: it’s important that we keep up. Back to the question. What design issues crop up for elderly users? Eye-tracking shows a lot of crossover with how dyslexic users scan pages. A more important question is how we get older users to engage with new technologies like social networking sites.

Julie points out that there is a wide range of cognitive disabilities; tiredness, short-term memory. Panayiotis says that navigation is a key problem. Presenting a user with a list of choices can be disorienting. We need to understand how cognition works to make navigation usable.

Julie quotes Joe: We live in a post-guideline era. We’ve heard that message again and again today.

Time for questions. First question: JAWS is ruddy expensive; are there any alternatives or ways of getting around the cost issue?

Kath responds that running a website through a screen reader like any other browser is not really testing it. You don’t get the user experience that a real user brings.

The guy who asked the question isn’t really getting that point. He wants JAWS to be free “like any other browser.”

Jonathan gives Thunder, the new free screen reader, a plug but warns that it’s not great with JavaScript. As for a free version of JAWS from Freedom Scientific, don’t hold your breath.

Bim agrees that JAWS is expensive but points out that you can get a free version that will work in perpetuity for 40 minutes at a time.

Mike jumps in to emphasize the point: web developers should not use screen readers. You will just get distracted. Instead, stop and think about how people use the Web. Once you understand the barriers they might hit, you can start coming up with solutions. Nothing will help you test your website like having real users test it.

Kath tells a story from Legal and General where sIFR was used and tested with a screen reader by a developer. It worked well enough for the developer but actual screen reader users were being put off by finding these Flash movies in the page.

Another member of the audience echoes the advice: don’t test with a screen reader; get a screen reader user to test. He goes on to say that accessibility is a quality control issue. He reiterates the point that was made many times today: accessibility is a user experience issue and user testing is accessibility testing.

Next question. On the subject of square wheels, there are techniques and tips that are supposed to be helpful but actually are harmful. He mentions the abbr pattern and talks about what he “learned” from Steve Faulkner today… in other words, he’s swallowing the FUD. So where do we go to know what is and isn’t best practice?

Mike gives his site a plug. Mostly though, he says that it’s important to read stuff and think about it instead of just taking solutions at face value. Mike claims that the microformats community did things without thinking them through. What bollocks! He’s claiming that people should test things but he’s just repeating the FUD about the abbr pattern which isn’t based on real-world testing…

Ian jumps in to defend microformats because of the data portability they allow. He mentions hCard. Mike rebuffs him.

Right, that’s it… I can’t take any more. Give me that mic please, Julie. I repeat my points about:

1. Be specific! You have an issue with the abbr pattern not with microformats in general, and certainly not with hCard!
2. Do some real-world testing with real screen reader users. Practice what you preach. The BBC did this by getting Robin in (and he found no accessibility problems with hCalendar).
3. Who are you to decide what is and isn’t human-readable? I find 2001-02-03 to be more understable and internationalised than 01/02/03 or 02/01/03 or 01/03/02 and I’m a human, not a machine.

Phew! I manage to get the last word in before the man from Mozilla gives a little spiel about Firefox.

Now everyone is being thanked for participating. I’m given a bottle of something lovely and bubbly as a token of appreciation. Awww!

Thanks to Robin, Kath, Gwen and everyone who put the day together.

Update Mike has responded on his blog about the abbr pattern discussion. There are a few errors in there:

1. The title of Mike’s blog post refers to a non-existent datetime microformat. Fixed. The hCalendar microformat uses a combination of two design patterns: the abbr pattern and the datetime pattern.
2. Mike says that I fail to realise is that Steve Faulkner has a long track-record of basing his findings on thorough screen reader testing. This is not true. I am well aware of the sterling work done by Steve Faulkner and Gez Lemon. I would describe Steve as rel="muse".
3. MIke incorrectly states the microformats community accepted the justification ‘Most screen reader users do not change their default settings for abbreviations and by default, abbreviations are not expanded’. The microformats community has done no such thing. The community is working towards alternatives for the abbr design pattern which has acknowledged problems… but those problems are founded in the semantics of the pattern, not the accessibility of it.
4. Mike claims that I am unaware of the screen reader testing he has done on the abbr pattern. That isn’t true. I am not only aware of the testing, I am very grateful for it. My point was that throughout the day, we heard again and again that nothing beats testing real sites with real people. As much as I appreciate the great work of people like Stev, Gez and Mike, there needs to be some acknowledgement that there’s between testing a test case and testing a real document.
5. For the record, when I pointed out that most screen reader users don’t change their default settings, I wasn’t suggesting that therefore the accessibility concerns are unfounded… but I do think that they are often exaggerated.

Just in case this hasn’t been made clear enough, let me reiterate:

• The microformats community is very grateful for the continued collaboration of accessibility experts like Mike and Steve. What got me upset — and this is the kind of Redectio ad Absurdum that I was railing against in my keynote — was hearing how concerns about one part of one microformat so quickly got turned into microformats are inaccessible. We’ve spent years fighting blanket statements like JavaScript is inaccessible or Flash is inaccessible.
• The microformats community is actively working towards alternatives to the abbr design pattern for including datetime information. I know that it might not seem that way from the outside—we need to do better at communicating activity and progress.