Canadian election Web sites flunk standards test

Findings

An independent, nonpartisan review of Canadian political Web sites shows that all sites tested:

1. do not meet Web standards, meaning their underlying code is grammatically incorrect
2. probably work correctly in only one browser, Internet Explorer for Windows, even though not all Canadians use that browser
3. usually don’t bother identifying the language in which they are written (English or French)
4. are somewhat inaccessible to people with disabilities

Canadians rely on the Web for information from political parties. But some Canadians will be excluded from those sites because they use the “wrong” browser or have a disability that the Web developers have not bothered to take into account.

What’s the issue?

1. Web sites have an underlying form – the code that causes a browser or other device to display the page.
2. HTML is the code that gives the page its structure, identifying paragraphs, headings, images, and other elements. (HTML means hypertext markup language.)
3. CSS is the code that governs how the HTML appears in a browser or other device. (CSS means cascading stylesheets.)
4. There are rules for the way HTML and CSS must be written. If you comply with those rules, you have written valid code and your site meets Web standards.
5. Valid code is similar to grammatical correctness. If you write valid code on your Web site, browsers know exactly what you mean and do not have to guess when they display a Web site.
6. Separately, guidelines are available to make Web sites accessible to people with disabilities.
7. None of the political sites tested meets Web standards or follows the guidelines for Web accessibility – even though both are easy to do.

Why does it matter?

1. For most people visiting the sites, it actually does not really matter. But for the minority who need valid code, it means a lot.
2. Standard code means that any device can interpret the HTML and CSS – even devices the owner of the site has never heard of.
3. Most developers create sites that work fine in Internet Explorer for Windows. Many developers do not even bother to test their sites in other browsers. (Or they actively exclude other browsers from their sites.)
4. However, standard code tends to work well (in most cases, in almost exactly the same way) with a wide range of browsers, including Internet Explorer and Safari on Macintosh; Mozilla, Firefox, and Opera on Macintosh, Windows, and Linux; and text-only browsers like Lynx, which runs mostly on Unix.
5. With standard code, in principle everybody who visits the site can enjoy and use the site equally no matter what browser they’re using.
6. In practice, there are small incompatibilities among browsers, most of which, ironically enough, are found in Internet Explorer for Windows. Nonetheless, standard code levels the playing field for sites, visitors, and the devices visitors use.

Who suffers?

1. Nonstandard code is optimized only for the browsers on which it was tested. That almost always means Internet Explorer for Windows.
2. But the majority of adult Canadians are eligible to vote, and most of them are online. They don’t all use one browser or operating system. (This may be news to you, but the Web is not the same as Internet Explorer for Windows.)
3. If you don’t happen to use the browser that is preferred by the site owner, you may not be able to use the site.
4. People with disabilities rely on standard code to make it convenient to use the Web – or even simply possible.

Accessibility

1. Disabled people use the Internet.
2. Sometimes they need no modifications whatsoever, but other times they do. For people who do require modifications, the way a site is written makes a difference. It’s important to follow Web standards and accessibility guidelines.
3. A blind person, for example, may use a screen reader – a type of software that sits on top of a browser and turns text, icons, and other information into spoken words or Braille or both.
4. A low-vision person may use screen magnifiers to make the displayed image big enough to see and read. Other disability groups use different adaptations.
5. These adaptive technologies rely on valid code. Research shows that noncompliant sites take significantly longer to use for disabled people.
6. In some cases, noncompliant sites can be so difficult to use that they are essentially inaccessible.

What tests did we run?

1. On May 28, 2004, my colleague Craig Saila studied the Web sites for the Liberals, Conservatives, Bloc, NDP, and Greens. He ran the HTML and CSS for each of these sites through the validators that test for grammatically-correct code.
2. All sites tested had errors. Not one site passed any of our validation tests.
3. I also did a manual check for basic accessibility issues. All the sites flunked those tests, usually in simple ways, like failing to provide a written description of images (an alt text). None of the errors could be considered make-or-break; it was still possible to understand most of each site, though with difficulty.
4. However, one serious error involves multimedia. Only the Bloc offered videos with captions for deaf viewers. No site offered videos with audio descriptions for blind viewers. (Videos on the Conservative site were not tested; they could not be loaded due to invalid page coding.)
5. We also looked at language: Did the sites declare the language they were written in? (Adaptive technologies need that information. It’s also required by accessibility guidelines.) Only the English-language Green Party site met that criterion.
6. Finally, we did a check of the underlying code to determine how the sites laid out the information on the page. The old way, using HTML tables, can be perfectly valid, but is not preferred because of its complexity (including the added work required to read table-based layouts in screen readers). The new way, using CSS to position elements on the page, is preferred. Only the NDP sites had CSS-based layouts.

What does this mean in the real world?

1. Political sites are all inaccessible to people with disabilities to some degree.
2. The sites use outdated, incorrect code. Their designers are using 1998-era techniques for a 2004-era Web.
3. Web developers seem to have created their sites for Internet Explorer for Windows and no other device.
4. The sites don’t handle the official languages well.

Summary of results

Here is an explanation of the terms used in the following table. Remember, HTML is the structure of the page and CSS controls the appearance.

DOCTYPE

In order to have valid code, you have to state which variation of HTML you are using. This document type (DOCTYPE) is either present or absent. It’s a mistake if it’s absent.

Character encoding

Characters – like letters, numbers, and symbols – have to be stored and transmitted according to certain codes. Did the site declare which code it’s using? And if so, was that declaration actually correct?

Language

Did the site declare, in its HTML, what language the site was written in (English or French)? Accessibility guidelines require that declaration.

HTML errors

Number of grammatical errors in the HTML. Any number other than zero means you have invalid code. Higher numbers are worse.

CSS errors

Number of grammatical errors in the CSS. Any number other than zero means you have invalid code. Higher numbers are worse.

CSS-based layout

Does the site use tables (the old way – inconvenient to screen-reader users and not preferred) or the newer and better CSS layouts?

Note that the following results refer to homepages only and not to other pages inside the site. If a site gave a splash screen (e.g., to choose English or French language), we selected a language and tested that subsequent screen.

About this test

This test was done as a free public service to evaluate whether or not political sites are actually working with 21st-century design principles.

We may update the results once or twice before the election on June 28, 2004.

Limitations

1. Craig Saila tested the sites on Windows; I tested them on Macintosh. We used a range of browsers, including Internet Explorer on both platforms, Safari, Opera, Mozilla, and Lynx. We do not claim to have tested every page, or even significant portions of a site, in every browser. We didn’t need to: We found problems in the range we did use, reinforcing our point about the need for standards compliance.
2. We tested only the most basic accessibility issues, including alt texts on images and captions and descriptions for video. If sites couldn’t get those right, they flunked the accessibility guidelines immediately; it wasn’t necessary to test further.
3. Circumstances could change: Sites could improve or worsen in standards compliance. The number of validation errors could change every day or every hour.
4. Site owners did not know we were doing the survey. Had they known, they might have improved their sites.
5. Captioning and audio description of online video are somewhat difficult, though the other issues we raised are not.

Is this page valid?

This cobbler’s children do not go unshod: This page and its stylesheet are both valid. If we missed something, though, do let us know.

About the author

Toronto journalist, author, and accessibility consultant Joe Clark goes back over 20 years in accessibility. He wrote the book Building Accessible Websites (New Riders, 2003) and nearly 400 articles for magazines and newspapers. Dubbed “the king of closed captions” by the Atlantic Monthly, he works on paid academic and volunteer projects and does paid accessibility consulting for clients, mostly in the broadcasting field.

He also founded Webstandards.TO, a social club for standards-compliant Web developers in the Toronto area that, curiously enough, lacks a real Web site. We do have monthly boozeups, if anyone would like to come along. Check our mailing list.

He is not a member of any political party; he isn’t affiliated with any political party; no money did or will change hands in this evaluation; and he most assuredly is not the Rt. Hon. Joe Clark.

Links

Web standards

• To validate a Web site’s HTML, visit the HTML validator, provided by the W3C. The W3C is the World Wide Web Consortium, a nonprofit group that publishes standards. The W3C published the HTML and CSS specifications, for example.
• To validate a Web site’s CSS, visit the CSS validator, also by the W3C.
• The Web Standards Project is a group that promotes the topic.
• Jeffrey Zeldman’s Designing with Web Standards is the definitive (even standard) reference on the topic.

Accessibility

• The World Wide Web Consortium (mentioned above) runs the Web Accessibility Initiative, which publishes accessibility guidelines. They’re pretty technical.
• We tested against the Web Content Accessibility Guidelines 1.0. We did not do an exhaustive test; this was a volunteer project, after all.
• The research mentioned above (“noncompliant sites take significantly longer to use for disabled people”) is from the Disability Research Council (U.K.). Their report is officially available as a PDF; I have an HTML version on my site. The report notes that “both blind users and non-impaired users took far longer on low-accessibility sites than on high-accessibility sites, and that this effect was not much more pronounced for disabled users: 51% longer for blind users, and 46% for non-disabled users. It follows that all users, not just disabled people, would benefit greatly from the measures required to make sites accessible and usable by blind people.”

Author’s and contributor’s sites

• Joe Clark (accessibility section)
• Craig Saila (Weblog)

Contact

If you’re the Webmaster of one of these sites, we will happily link to your comments in response to this survey.

One more time

The author of this survey, Joe Clark, is not the Rt. Hon. Joe Clark.

Revision history

2004.06.03

Posted.

2004.06.20

Copy errors fixed.