Web Accessibility refers to making Web pages usable to people with disabilities. Probably the most significant example of a disability that makes it hard to use the Web is blindness or poor vision. (When I say most significant, I'm speaking in terms of having the biggest problems in normal Web browsing software that can be avoided with different software.)

There is a significant community with the goal of improving the accessibility of the Web to users with disabilities. However, I think this community is in significant danger of being taken over by, or at least best known by, those within it who espouse such extreme positions that they risk causing the entire community to be ignored.

In a recent message, John Foliot expressed a desired to have “either stopped BeSpin in its tracks, or challenged its creators to solve the problem before launching the application.” He said, in other words, that Bespin, which is an experiment in its early stages of development, should not have been allowed to be put on the Web until it was made accessible. I think this argument ignores the basic idea of how invention works—through lots of experiments—and would, if enforced, suppress lots of beneficial invention and innovation. Would he argue that the telephone should not have been allowed to be sold in the United States between its invention in the late 19th century and its accessibility to the deaf in 1973? Even if he wouldn't, I think his argument would, which makes it a bad argument.

Another set of extreme positions I've seen relate to refusal to recognize the widely varying nature of content on the Web. Existing disability law, for example, might require installation of wheelchair ramps in places of “public accommodation,” but doesn't require them to be installed in everybody's houses. Likewise, I would expect an online form on a government Web site that is required to visit the United States to be usable by blind people, and likewise expect good alternative text for an image on a government Web site describing how a bill becomes law. Yet I would not expect a bunch of photos that John Smith shares with a few friends on a Web site to be required to have reasonable alternative text. In other words, content on the Web varies widely in importance and amount of use. Yet some accessibility advocates insist that even John Smith posting a few photos online must be forced to provide equivalent alternative text to replace the photos.

I've also seen a number of arguments, such as one from Matt Morgan-May and another from John Foliot fall back on arguing that Web accessibility is a “human right.” I'm suspicious of any arguments that invoke human rights, as doing so is essentially playing a “get out of jail free” card. There are very good reasons for many of the things that are commonly referred to as human rights. Those arguing for one should instead lay out the underlying reasons. For many of the existing concepts people refer to as human rights, this is a relatively straightforward argument. However, Web accessibility involves tradeoffs, such as between burdens on those who send information and burdens on those who receive it. Sensible choices along this spectrum can vary depending on how the Web is being used; there's a big difference between publishing to an audience of five (that might be larger later, if you happen to succeed) and publishing to an audience already known to be in the millions.

I think these extremist positions tend to orient much of the accessibility community towards casting blame on evildoers (such as the evil John Smith uploading his photos to flickr) rather than solving the real problems of disabled users. It's far better for accessibility to be the automatic result of writing HTML in the normal way rather than something that has to be done as an extra development step, as Henri Sivonen described. But I think the attitude that evil Web authors need to be forced to care about accessibility leads to technically worse solutions that require more work for authors and leave the Web less accessible to disabled users as a result.