The bullhorns and short-term thinkingWednesday, June 01, 2005
The major participants in the accessibility community are well-intentioned and good-hearted. We all want to make the world a better, more equitable place - and in our case we want to make the Web part of the solution. Yet, we seem to be fighting each other, much like Monty Python's Life of Brian where multiple organisations are fighting for the right to free Judea from the Romans.
I suppose its the conflict that creates a great deal of uncertainty. There's more than one right answer, yet there is tension when our right answer isn't the preferred option. And its occurred to me that one of the reasons for the tension is that we are missing an important party in these discussions.
Disability organisations: RNIB
The RNIB have been leading the advocacy of accessible websites - amongst their other endeavours. Their recongition is deserved, they've fought exceedingly hard to protect their members from alienation.
When it comes to assistive technologies, they know their supported products inside out. They are strongly aware of what technologies work today, and what worked yesterday - and they can explain it to you.
Two years ago, when the RNIB had their website relaunch disaster, I lost a lot of faith in their abilities to guide the web development community. Their site was, and still is, accessible to their core audience. But as a practical example of technically proficient accessible web development, it left a lot to be desired.
One of the main criticisms I have of the RNIB is that because they are so dedicated to their members, there's a tendancy to disregard things like modern web development techniques, and keep using out-dated techniques - tables and one-pixel image school of web design. What made it worse was actually defending those out-of-date practices.
In their defence, they saw their development techniques did produce an accessible website, and they correctly pointed out the deficiencies of my attempted improvement. I think, I, as a web developer failed to make the point clearly.
Over the past two months I've seen a different, and more encouraging, side of the RNIB - particularly their consultancy wing. Experts like Donna Smillie and Bim Egan, who demonstrate an appreciation of modern web design techniques, understand the accessibility problems, and understand it enough to explain to others, including myself, in a clear and concise manner.
Developers and designers of accessible websites
We build sites, and we take a lot of pride in them. The technical accessibility experts spend a good deal of time perfecting their techniques. We buy into the idea of web standards, well structured markup and stylesheets for layout. We understand both the short term and long term benefits of supporting webstandards. Yet, we struggle to get a clear message across.
One of the most difficult arguments is trying to explain why our modern development techniques are better for accessibility than the 1995 era tables and the one-pixel gif approach. How do you prove the case?
Its not enough to point to the Web Content Accessibility Guidelines, particularly checkpoints 3.3, 11.1, and 5.3 and declare victory. The obvious retort is that these checkpoints don't improve the accessibility of websites to screen reader users right now - the evidence is there, blind people have been participating on the web for as long as it has been around, which is at least a decade before web standards became more than just a good idea.
To the RNIB, we web developers probably look and sound like a group of people with bullhorns, shouting our message over and over while the rest of the world walks by in morbid fascination. Since our message sounds like gibberish, and bullhorns don't make it sound any better, its noise for the sake of noise. Look at me! Look at me!
We've missed the real argument here. Screen readers in use today don't take much advantage of structured and semantic markup. The reason is obvious, 99.7% of websites are not built using properly structured markup - they are built with the "it works in Internet Explorer" philosophy - with various definitions of work.
Screen readers, unsurprisingly, are developed to handle real-world websites. So they are not optimised to take advantage of correctly used headers, not optimised to drop straight into tables navigation mode when a table is encountered. Lists are treated the same way as multiple paragraphs - hardly an indication that a list is even present. A typical screen reader ignores most of the common markup structures. Why? Because they are not being used correctly by website builders. The IT profession has a wonderful acronym GIGO - garbage in garbage out. If there was a "use the structured markup" option on a screen reader, 99.7% of sites would immediately be more difficult to use.
If all websites used properly structured markup, accessibility would be better, since screen readers would have a perfect excuse to leverage up the markup of a page to the benefit of screen reader users. But we are nowhere near that sort of buy-in power, and until then the benefits of checkpoints 3.3, 11.1 and 5.3 remain a pipe-dream.
Man in the middle
So the RNIB point out real life situations, and what works now. Web developers point to specifications and talk about sacriligious twisting of web standards. Both sides shake their heads in disbelief, muttering "get a dose of the real world".
We are missing an important group of people. The ones who sweat blood to produce assistive technologies like screen readers. We've somehow managed to leave them out of the conversation. And they have an important role to play.
They can translate the RNIB's position into something web developers can understand: we are stuck with too many hideous websites that we have to be able to access. The techniques we are teaching work in the current range of screen readers. Today, screen readers don't, and probably can't take advantage of well structured markup, because there's so little of it out there in the real world.
They can also translate the web developers position into something more meaningful: Yes, the quality of markup is terrible today, and its preventing us from delivering a better experience to screen reader users. We need to encourage better use of structured markup, and stylesheets for layout. In the long term, screen readers can start supporting such clean markup, and start trusting that web developers are using markup for the purposes it was intended.
Web developers are not actively looking for ways to avoid creating accessible websites. We are looking for ways to improve the shockingly low levels of compliance today. We see that the 1995 web development techniques, in the long-run, are a dead end. We see that the future of highly accessible websites lie in the adoption and understanding of web standards. At the end of the day, we want all people, whether using assistive technologies or not, to use our sites.
Delegation of responsibility
The screen reader developers need to be more involved in the conversation. We web developers need their support and commitment to web standards. Web developers alone cannot deliver on the promise of web standards - we rely on screen reader developers to support those standards too. And we realise that our quality of markup has to improve, and we are reliant on screen reader developers to render that structured markup to the end-user as best they can.