Web Accessibility briefing

Introduction and context

The Disability Discrimination Act makes provisions about the accessibility of web pages. This has been coming for some years, presaged by legislation in the USA and other countries, and warning shots fired by the Olympics case in Australia, various court Opinions in the USA, and so on.

Fortunately, the original intentions of the WWW were to make content available to a wide range of different browsing situations, and this works very much to the WWW's advantage: compared with, say, buildings, or means of transport, for which the provision of disability access can incur considerable additional costs, the WWW behaves quite differently: basic accessibility techniques come as part of the method, and many of the options are beneficial not just to a few individuals, but can bring benefits to many WWW users whose browsing situation is a little different from the mainstream.

W3C have been working for some years on web accessibility, and some excellent resources have been prepared to assist all those concerned with web design (web page authoring, client agent development, and so forth). Sadly, we have to note that there seems to be entrenched support, maybe promoted by those with a professional background in print design, for desktop-publishing (DTP) based solutions, and so-called "WYSIWYG" web page tools, whose proponents purpose to deliver the same pixel-exact visual presentation to every reader, whether they can comfortably use it or not; and who mischievously claim that accessibility is impossibly hard and requiring unreasonable investment of time and effort on top of their original design work: thus defeating the benefits which the techniques of flexible design can bring or, in short, to defeat the very benefits for which the WWW was devised.

Now, whereas it is true that the development of the WWW in the past has been somewhat patchy in this regard, we are fortunate that many things have been falling into place in the last few years to make the plan a practical one, as I hope to show. And browsing situations have been getting more diverse with time, with some users getting higher-end desktops while others use small hand-held devices, which all points to the advantages of a flexible approach to content delivery - which also benefits disability access.

Modern browsers, such as Mozilla (Netscape 6/7), Opera, MSIE, have entirely confirmed the benefits of the original plan; and web pages which are constructed in accordance with the principles are basically accessible to a wide range of browsers, and to a wide range of browsing situations: this already brings considerable benefits for accessibility, while at the same time potentially simplifying the task of web page authoring, if it is approached in an appropriate way.

What is disability in this sense?

The most immediate disability which comes to mind in relation to web page access is blindness or partial sight. But this is not the only issue. Some users have problems with physical co-ordination, causing difficulty in using a mouse (think: imagemaps); if audio media are offered, then deaf users would be at a disadvantage, indicating (if the audio contains significant content) that a transcript should also be made available.

There are other issues: for example, anomalies of colour vision are far more prevalent than is often supposed, especially among males. While it's by all means a good idea to utilise colours as an auxiliary marker of content, in a WWW context it should not be relied upon as the sole indicator of some critical distinction: use other indicators to support the message.

As remarked before, the provision of accessible content is beneficial not only for readers with disabilities, but also for normally-able readers who are in difficult or unusual browsing situations. Not to forget the indexing robots, without whose co-operation a web page is adrift on the WWW without hope of rescue. As another commentator put it:

The most important visitors to your page are blind and deaf, and use a browser which you have never met.

What approaches are there for making a web site accessible?

Looking at what is happening 'out there', I think we can crudely categorise several kinds of approach: in reality, usually some mixture of these approaches is taken. 'Bobby' is a checker of objective accessibility features, of which, more later.

Review the existing legacy pages; if found inadequate, then take the opportunity of re-engineering them to current best-practice.
"Take legacy pages, and bludgeon them until they pass Bobby"
"Make a separate text-only site" (or use an automatic processor such as BETSIE)

I guess it's no secret that I favour the first option in principle, albeit there may be compromises to be made in any of the approaches. More about these compromises later. The first approach may sound like a lot of work; but if the content is due for revision anyway, then "doing the job properly" in a systematic way may very well take less effort in the long run than trying to bludgeon unsuitable material piecemeal into shape.

Major issues

Issue:	First, the bad news:	Now, the good news:
Status quo:	The majority of pages on today's WWW fail both objective and subjective accessibility criteria, even at the least exacting test level; ours are generally no exception.	There are recognised repair techniques and tools. Tools are available (Bobby; A-Prompt; Accessibility Valet) for pointing up areas of concern; however, although some problems can be repaired "by rote", others call for a more considered approach, and the one immediately suggested by the tool might be a very suboptimal choice! - Bobby's prominent suggestion to 'supply an alternative text-only site' is almost never the best choice, and should be regarded as very much a last recourse.
Design:	"Presentational HTML", as exemplified by HTML3.2, reached the end of the road some time back, but its proponents are still very much at large, and an obstacle to real progress. Many commercial web page designers have a background in visual design and paper publishing, and their favoured design tools are unfortunately strongly influenced by this.	Techniques of flexible design not only produce better results across a wide range of browsing situations, offering basic accessibility support as a matter of course, but can be applied without impairing visual results in mainstream browsing situations; and furthermore are more maintainable since, by and large, only the content structure is reflected in the HTML source, whereas the presentation is delegated to a stylesheet, which can be developed by a specialist, and used across a whole corpus of content.
Legislation:	The DDA required the necessary steps to be taken "by Sept 2002": a University flyer informing this department of the requirement reached the webmaster 3 weeks into September 2002.	(A.F has taken a personal interest in this issue since before the W3C WAI was started, and has been urging colleagues, though at first with limited success, to design web pages accordingly.)
"WYSIWYG":	*"What you see" is not* what others get.** Attempts to achieve pixel-exact design in HTML are not only unsuccessful in the wider scope of the WWW, but also harmful, resulting in web pages heavily overburdened with "presentational markup" that's counter-productive outside of the narrow range of presentation for which they were designed; and which carry a long-term legacy in terms of page maintenance and which inhibits 'content re-use'.	The original intention of HTML as predominantly marking-up the logical structure of content, with separate stylesheet(s) to suggest appropriate presentation, has been vindicated in HTML4 with CSS2, is rather well supported in recent browsers, and can be used in ways which are not harmful to older browsers and to less-usual browsing situations, albeit the presentation in older browsers might be rather plain.
Web page content:	Textual content is considered to be the most widely accessible form of content, thanks to the ability to convert it into spoken form, Braille etc.	Accessibility by no means deprecates the use of other media for appropriate purposes: the widely-heard criticism that 'HTML purists' want all web pages to consist of nothing but black text on a mid-grey background is - and has always been - very much a 'straw man argument'. Accessibility guidelines, however, call for textual alternatives to be offered for such content.
Non-text content:	Accessibility calls for meaningful textual equivalents for non-text content (images, sounds, interactive displays etc.). *"Meaningful" equivalents does not* mean texts like "here is a picture which you can't see" or "this is for layout only",** as is, sadly, too often seen in would-be accessibility fixups from the DTP crowd.	Accessibility is achieved by using other media in ways which do not bar access to the content for users with disabilities. This means on the one hand offering textual alternatives to media which are not in themselves accessible; and on the other hand avoiding hurdles of inaccessible navigation (such as frames, mandatory javascript, imagemaps needing fine physical co-ordination to operate, etc.) which bar access to content which would, in itself, be accessible.

Separation of Content and Presentation

Some authors are heard to complain that their web pages are conceived as a visual entity, so they cannot separate content from presentation. Well, the WWW's answer to that is that the WWW inevitably does separate content from presentation, when their web pages are served out to a widely diverse range of presentation situations. So if it's going to happen anyway, why not take a moment to understand the idea, and see how to capitalise on its strengths, rather than bemoaning its weaknesses?

So how to achieve this wonder, you ask?

This is where one individual cannot solve all problems! From a technical point of view, I can only point out what the issues are, and suggest some considerations which could help.

I have to say that I've found the University's central response to be rather disappointing, and probably not a good example to follow verbatim, although it does contain useful material too.

One point that comes out of their presentation however is that they recommend DreamWeaver. My assessment of that, on the basis of limited knowledge, is that it's probably the least harmful of the widely-available visual styling tools, but it does need to be used with an appropriate knowledge and understanding of the underlying structure.

The department evidently moves towards having the content providers composing and updating their own web pages, rather than delegating the task to a specialist. And, understandably, authors have used whichever tools they had to hand, no matter what their suitability in regard to accessibility. And as was mentioned before, most of the obvious tools have been designed on the (inappropriate) presumption that making web pages is solely an exercise in point-and-click visual styling, without any particular attention to the underlying structure.

I don't know a single complete answer, but I do see quite a number of elements which could be applied positively in reaching a viable solution, and which would be beneficial in their own right, rather than being a lot of extra work for the benefit of a small minority.

The recent updates to our main public pages have, I feel, been a successful co-production between someone with an eye for visual styling (i.e Johannes), and another with a somewhat stubborn and pedantic attention to structural detail (i.e myself). And in fact these pages not only pass most or all of the objective accessibility tests, they also seem to me to fulfil the subjective requirements too. However, these represent a relatively small number of presentation pages for our web visitors, rather than the day to day bread-and-butter teaching materials which staff are providing for their students, and which also fall in the scope of the DDA requirement, as indeed do all web pages offered in the name of the Department: the legal position is that we need this irrespective of whether we currently have any students in the class who require it.

[Aside: the status of users' personal web pages, i.e those with /~username/ in their URLs, is unclear. The University's note implied that all web pages on University servers were required to comply. The department normally leaves the responsibility for personal web pages with the page owners themselves, within reasonable limits of legality and decency. Naturally we would urge personal page owners to provide accessible pages, irrespective of whether the department is supposed to be trying to enforce it.]