Worth a thousand words?

Graphics add pizazz to the World Wide Web – but who is being left behind in the graphics frenzy?

Note: This article, dating from the dark prehistory of the Web, appeared in the Toronto Globe and Mail newspaper as a form of provocation. Unlike every other accessibility advocate on earth, I am not actually opposed to graphics. I go back 20 years in graphic design, and I have some very stern words indeed for anti-graphics accessibility advocates in my forthcoming book. Consider this article an historical artifact.

Try to explain the appeal of the Internet's World Wide Web to net-ignorant friends and you're apt to mention the fun and impressive graphics available online. In fact, you're likely to mention those jazzy graphics well before you even attempt to explain the concept of hypertext, if you even get around to mentioning that capacity to zip from one chunk of text to a related chunk anywhere else in the world and thence to another chunk, and so on. In its brief history, the Web has been transformed from an experimental means of presenting multilayered information to a place where Webmasters can show off their ability to scan in photos and add groovy graphics.

That's fine and dandy if you have a cream-of-the-crop modem and the SLIP or PPP connection usually necessary to run a graphical Web browser program like Netscape Navigator, but not everyone is so lucky. Internet critics often issue stern warnings that the net's explosive growth threatens to divide the world into information haves and have-nots, the latter group being entirely shut out of the information age.

But that polarized view is simplistic. There still exists a broad range of technological sort-of-haves who use outdated computers (aging Mac Plus; dusty 286; discontinued Atari or Amiga) and old-style text or character-mode Internet connections. In addition, thousands of blind and visually-impaired people are online. For those groups, the currently fashionable reliance on graphics to actually communicate on the Web amounts to a barrier. Are those netters on the verge of being shut out of the flashier forms of Internet expansion?

If you have a text-only (or shell) account – which requires fewer system and modem resources than SLIP or PPP connections, hence are the norm for students at cash-starved universities – you can explore the Web using Lynx, a widely-used shareware text-only Web browser originally developed by the University of Kansas. While unsexy, Lynx is fast: With no bloated graphics to download, Web pages appear in mere seconds. Quick response times made the locution “surfing the Web” a reality.

If every graphic image were purely decorative (like corporate logos), Lynx users wouldn't be missing much, but hot-to-trot Webmasters are increasingly whipping up graphical buttons and clickable "imagemaps" for hypertext links that really ought to be straight text. With Lynx, these typically appear as [IMAGE] or [LINK] – not very helpful.

With each graphical image, Webmasters are supposed to include an alt definition – a short pithy text-only description of the graphic. If you’re using a graphical browser like Netscape and have “Auto load images” turned off, Netscape will refrain from loading graphics and will show you instead a generic icon meaning “graphic” along with the text of the alt definition. (Running Netscape with graphics off is the easiest solution to a prime complaint of Web-surfers: slow speed.)

But Larry Jackson of the National Center for Supercomputing Applications (NCSA, home of the original Mosaic browser) at the University of Illinois points out that “[Web] authors seldom even read, much less implement, standards-compliant documents. Almost none of the pages we review fully pass the syntax checkers. There’s a sensitivity problem involving the pool of [Web] authors that is going to be tough to solve.”

The burgeoning Free-Net movement in the United States, Canada, and elsewhere – in which community-based organizations pull strings and round up donations to provide free or nearly-free net access to anyone who wants it – is getting a kick in the virtual knees by these graphics-heavy sites. Tom Grundner, president of the National Public Telecomputing Network in Cleveland (which owns the term Free-Net; now defunct), estimates that "maybe a third or more of our systems have Web access and that would be primarily with Lynx."

As Michaël van Eeden of the Digital City project in Amsterdam says, “We especially cater for those people who don’t have the money for an account with a provider, so we have to think about text-only access. We provide dial-in, where you can choose [from] PPP, SLIP, and Lynx, and most people still use Lynx. We try to make extensive use of the alt tag, but sometimes you’re just not able to. More and more information providers on our system want to use clickable images, just because it looks cool.”

And consider blind and visually-impaired netters. Even approximate numbers of blind people online are impossible to come by, but blind people have long been users of computers. Some with modest visual impairments can use ordinary computers with software that magnifies the onscreen image. Those with little or no usable vision tend to rely on screen readers, a generic term for software/hardware combos that selectively read the contents of a display out loud using speech synthesis – not only text, but menus, buttons, and other parts of the graphical interfaces sighted people take for granted. What does a screen-reader user do in the face of graphical Web pages?

The obvious approach would be to use a screen reader with Lynx, since Lynx is itself optimized for text-only work. But that's not what Josh Miele of Berkeley Systems in California does. Apart from selling flying-toaster screen-savers, Berkeley also produces a screen reader called OutSpoken for Macintosh and Windows [since sold to Alva Access Group]. Miele, a product developer in the Berkeley Access division who is blind, runs Netscape on his Mac with graphics turned off. “If I want the graphic, I can always download it,” he says. “But chances are I don’t want it and it would just slow down my connection and put garbage on my screen.”

Lynx is passé, Miele notes, because graphical browsers “are the ones that are going to be updated. Those are the ones that are going to have the latest features in terms of interpretation of forms and so on.” OutSpoken does automatically read aloud the alt text if it exists, and the filename of the graphic if desired (filenames are often quite descriptive); you can assign a name to a graphic image you run across frequently that OutSpoken will use thereafter, and the program will speak that name in a special “graphics voice” if you so choose.

Still, those are workarounds. Though the popularization of the Web is still only beginning, it may already be too late to expect Web developers to use graphics judiciously, not promiscuously, and take the time to ensure that their Web pages are accessible to all, even via the dirt-simple technique of alt tags. Failing that, the Web may continue its devolution into a fancy-schmancy photo gallery optimized for a techno-elite.