You are here: Home / Dog Blog / Archives

Ditching Accessibility?

Tuesday 2 March, 2004 ( 5:27PM GMT)

I have decided to stop claiming that HTML Dog is 'AAA' compliant, the highest level of accessibility as defined by the W3C's Web Accessibility Initiative.

Unfortunately, I think that claims of WAI guideline compliance often follow an attitude of 'ah, well, it's close enough' or are simply based on something such as 'Bobby AAA approved', which is quite meaningless because there are so many factors that an automated program cannot take into account, which can leave a completely inaccessible page that falls way short of the intended WAI AAA.

It isn't as easy to judge standards compliance when it comes to accessibility as it is when it comes to code, because there is so much that is open to interpretation. But in my opinion, even after any interpretation, sites claiming compliance often bend the guidelines to such an extent that they do not live up to their claim. A 'near enough' attitude is like having a perfectly valid XHTML 1.1 document bar a few font tags here and there. It is simply incorrect to claim that you are compliant with something if even the smallest detail of a standard is not adhered to. You can't pick and choose - that's the whole point of Standards.

'Guidelines'?

A possible train of thought regarding the Web Content Accessibility Guidelines is "OK, these are guidelines. I have considered them all, but decided not to implement a single one. I can still call my web page AAA". Logically, for guidelines, a pedant could claim that this argument is quite sound. As guidelines, the WAI recommendations are great and this is where they hold their strongest value but when the 'A', 'AA' and 'AAA' stamps are used, this becomes a suggestion that the guidelines have been accepted and adhered to as rules.

I'm not sure how possible it actually is to achieve true unarguable AAA compliance without making some design compromises. As one example, I prefer elastic elements, but absolute units are sometimes required to achieve a certain effect. HTML Dog uses one-pixel width borders and the header now has a fixed height measured in pixels. WCAG checkpoint 3.4 reads "Use relative rather than absolute units in markup language attribute values and style sheet property values." - no mention of text - units in general. So according to this, even using pixel-width borders is not compliant.

But these accessibility 'compromises' I have made are not compromises at all - the border widths are insignificant and the graphical header is sufficiently large enough for the visually impaired to read and actually not overly-important to the actual content anyway. It is my argument that the site is just as accessible.

Another area where I would say that HTML Dog also breaks WCAG is when it comes to accesskeys (many of which have been removed due to conflicts with browser shortcuts).

So What To Do?

Many might not argue with a statement that HTML Dog is 'AAA compliant' because the WAI is often vague and can be interpreted in different ways (often by tempting a few round pegs into square holes). In fact I think that this site is more compliant than many web sites that claim this high accessibility standard. But I am simply no longer comfortable with either strictly adhering to the WAI guidelines or with bending them and claiming to adhere to them.

I believe that the WAI. is a good thing and that standardising accessibility initiatives is a necessary step towards their proliferation. Ahead of the WCAG 2.0, one solution to this problem could be to have an independent accessibility statement built around the WCAG.

The WCAG 2.0 may well clear things up and be more up to date but at the moment we're stuck with version 1.0 and although it may be flawed, if a web page claims to adhere to that standard then it should.

Comments

Comment 1

This makes me happy.
Someone is standing up to the W3C, PTG of all people. I thought if the W3C said all web developers should jump off bridges he would. Wow. No offense.

Do you think that Microsoft should follow the W3C event model and the Box Model, even if they make no sense?
To quote Mr. Koch:
"No doubt the Microsoft model is better than W3C’s. 0 should mean 'no button pressed', anything else is illogical."
...
"In my opinion W3C has made some serious mistakes in defining button. Fortunately only Mozilla follows the spec, all other browsers have opted for the saner solution."

I just wanna know what you think.

So said Dante Evans on Tuesday 2 March, 2004 at 11:37PM GMT.

Comment 2

I know what you mean about parts of the event model but the W3C's box model is absolutely fine and logical. No one has ever disputed that.

Anyway, all that is completely off the point of the post. I think at the moment web developers (and clients) need a label on their site to say "We have thought about accessiblity". WAI tags are being used to perform this function, which is wrong but understandable.

I've put a AAA tag on my site, and yes, I argue that validity of some of the guidelines for my particular site and content. My normal process for validating a site for accessiblity is to check through the WAI guidelines. If I ahere to a guideline thats fine, if I don't then I'll make sure I take measures to get around the problem one way or another. Like you say, adding pixel measured borders does not affect accessiblity. There are many examples of these kind of loopholes and grey areas in the guidelines.

If you keep in mind the objectives of the WAI guidelines rather than interpreting them word for word I think it's a very useful thing. However, too many people don't really understand the essence of accessibility and just bowl through their site with Bobby. This is not useful.

So said Dan W on Wednesday 3 March, 2004 at 1:01PM GMT.

Comment 3

Dan, you don't HAVE to agree with everything PTG says just because you wrote an article with him.
Jk, but you're wrong:
http://www.quirksmode.org/css/box.html#link2
Explain to me how the box model makes sense. I want to know what the pro-standards people think.
I don't display "Valid CSS" or "Valid XHTML" stickers on my site. In a real website that is visited by people with no knowledge of HTML (my site is such), it's completely useless and confusing. I try to make it accessible, but I don't think the W3C should have guidelines for EVERYTHING! Jeezz.

So said Dante Evans on Thursday 4 March, 2004 at 12:22AM GMT.

Comment 4

"Explain to me how the box model makes sense."

How would you apply borders or padding to a fixed-size image? Or the letter 'M' ?

So said Hemebond on Thursday 4 March, 2004 at 12:34AM GMT.

Comment 5

That doesn't make any sense. You wouldn't apply borders or padding to a fixed-size image; if you did it wouldn't be fixed size. The letter M? What the hell are you talking about?

So said Dante Evans on Thursday 4 March, 2004 at 1:33AM GMT.

Comment 6

Off topic people. Although I do like Hemebond's simple explanation of why the box model so obviously makes sense.

So said Patrick on Thursday 4 March, 2004 at 10:12AM GMT.

Comment 7

Getting back on topic again: you still could claim full AAA compliance and define your borders in pixels since px is defined as a relative unit in CSS (relative to the device). You just can't use pt or cm which are absolute (even though they are known to shrink over time or depending on the speed you move, but that's a different set of rules). Anyway, what does it show us? That the guidelines are, at times, simply braindead or at least unapplicable, but people still slap a AAA sticker on their pages.

So said Tomas on Wednesday 15 June, 2005 at 9:33PM GMT.

See Also

^ Top

SiteGround: Fast, reliable, recommended hosting.