Accessibility
Dear visitor,
This web site is no longer run by First Accessibility (Test Partners Ltd). A new site is coming shortly. In the meantime, the content of the old web site is available below.
Thank you for your patience.
accessibility.co.uk
By Steve Green, Technical Director, 10 June 2007
The WCAG Samurai are an independent group of developers convened in 2006 to formulate a set of errata (corrections) for the Web Content Accessibility Guidelines 1.0 (WCAG 1). These were delivered on 07 June 2007.
The Errata are published on a website at http://wcagsamurai.org/, which contains links to two independent peer reviews by Gian Sampson-Wild and Alastair Campbell. A final version of the Errata will be published after the peer reviews and other contributions have been considered. This is my contribution.
In most respects I agree with the contents of the Errata and comments by the peer reviewers, so the following comments reflect where I disagree with the reviewers or make additional points.
Whilst there is no doubt the WCAG need to be updated, the lack of W3C approval is likely to limit the commercial use of the Errata. The unofficial status coupled with the likely additional costs (especially if there is multimedia content) makes the Errata an unattractive proposition.
The ideal outcome is that the W3C adopt the Errata, which is perhaps what the Samurai had in mind from the start.
The Errata are virtually unworkable if they are not incorporated into the WCAG. In practice every designer, tester and consultant will incorporate the two documents themselves, the result being that no two versions of the merged documents will be the same. This is crazy and will inevitably lead to disagreement.
In addition to the reasons the reviewers have mentioned, the time (and hence cost) required for testing should be considered. If the Errata are combined into the WCAG they will take no longer to test than the WCAG do now. However, if they remain separate the time required to test them will increase.
The paragraphs at the end of the Errata suggest that there is a legal issue with combining the documents. It would be useful to have an explanation because presumably we as individuals would be subject to the same issue.
If the Errata are not incorporated into the WCAG then it is essential that they are numbered such that they can be unambiguously referenced in design specifications, test reports, correspondence and other documentation.
Each of the Errata should be cross-referenced against all of the relevant WCAG checkpoints. We can do that ourselves but we should not have to with a document that is being presented as a standard (albeit an unofficial one).
The removal of "Until user agents..." from the WCAG is most welcome, but the Errata have been based on the existence of Utopian user agents that fully meet the requirements of the User Agent Accessibility Guidelines.
The Errata should not make undue compromises to accommodate bad user agents, but nor is it reasonable to entirely ignore the reality that the deficiencies in the user agents currently in use are not going to get addressed any time soon. Whilst Internet Explorer is usually seen as the main culprit, Firefox shares many of the same deficiencies in terms of support for keyboard navigation, image maps and other criteria.
If we are interested in positive actual outcomes rather than theoretical levels of accessibility then we need to take account of the user agents that are actually available to users rather than those we wish were available.
The errata state they "must be made intrinsically accessible..." without defining what this really means. Flash accessibility features allow the tab sequence to be controlled and buttons to be labelled etc., but the biggest problem we encounter in user testing is that the user doesn't have a clue what is going on or what they are supposed to do.
The problem is cognitive (although not due to a cognitive disability) rather than technical. In fact an audio description might often provide the necessary information, although this is not suggested in the WCAG or the Errata.
I do not understand why these are permitted when far more stringent restrictions are applied elsewhere, particularly with respect to multimedia. The provision of 'alt' attributes does not mitigate the problems these graphics cause because they affect those who can see the graphics, not those who cannot.
Graphical representations of text can rarely be resized, and they often become less readable when they are. The colours they contain can only be changed if the user agent supports colour substitution, which few do. I believe they should not be permitted except for logos, unless they are duplicating a plain text equivalent.
The Errata no longer require that redundant text links are provided for client-side image maps since this is a Priority 3 checkpoint. Whilst this would be reasonable if user agents handled image maps appropriately, even the best ones do not.
I do not believe the Errata should ignore the fact that at in least 95% of user agents it is not possible to use an image map if images are not displayed, because the 'alt' attributes for the hotspots are not displayed so the hotspots cannot be identified. Redundant text links should be mandatory.
Under the WCAG 1.0 the colour contrast for text was a Priority 3 checkpoint. Since the Errata state that all Priority 3 checkpoints should be ignored, they should also state explicitly that adequate colour contrast for text is now a requirement. The existing reference is not sufficient.
Replacing the phrase 'sufficient contrast' with 'confusable colours' makes sense but it is not very helpful. How can a designer or tester know if two colours are confusable? The Errata appear to be saying that the WAI colour contrast algorithm is not a suitable means of assessing confusability, yet they do not suggest an alternative.
It is interesting that whereas all the other checkpoints are defined in terms of the measures that should be taken rather than the actual outcomes, this checkpoint is defined in terms of the outcome with no guidance as to how it can be achieved. If the Samurai don't know, how should anyone else?
I note that the two peer reviewers are split on this point, with Alastair saying "the guidelines on colour contrast are quite clear", and Gian saying "Some clarification would be useful here".
It is absolutely true that user agents should be able to resize text regardless of the units used to specify its size. However, it would be reckless to ignore the fact that at least 75% user agents cannot do so.
The rationale for changing this checkpoint has not been explained, and there is a suspicion that it is driven by dogma and a desire to punish people who choose to use Internet Explorer. If that's true then the Samurai should have the courage to say so. Otherwise they should explain the purpose of making changes that are of no benefit to designers and will greatly inconvenience three quarters of users if implemented.
I am content with the use of absolute units to specify the layout.
I would endorse Gian's comments regarding notification of processes that are subject to session timeouts. This is a Section 508 requirement and should have been a WCAG requirement from the outset. Although it may be considered to be a general usability issue, it impacts users with disabilities to a greater degree than others, and therefore warrants inclusion in the Errata.
The user should be informed in advance of the following:
Furthermore, the user should be explicitly informed when a timeout has occurred.
'title' attributes are not a suitable means by which to notify users if they are the sole means of notification. Not all user agents support them, and in all the major graphical user agents they are device-dependent. Maybe they shouldn't be, but they are.
It is not uncommon for form controls to be provided without any kind of label, yet labels are not mandated by the WCAG or the Errata. Usually the purpose of the form control is obvious to most users because of the context, such as a textbox being to the left of a Search button. However, the purpose of the form control may not be obvious to all users.
Similar problems occur with two-dimensional matrices of radio buttons, as are often found in questionnaires. Typically the radio buttons do not have individual labels, but their purpose can be determined by the headings at the top of the column and end of the row. Such an arrangement is invariably incomprehensible to a screen reader user and can be in other user agents if the content linearises rather than remaining in a matrix.
I would like to see the Errata state that implicit labels are mandatory, with the provision that they can be positioned off-screen as long as this does not reduce the comprehension of the form. Of course the implicit labels must be explicitly associated with the form control.
The Errata state "HTML variations developed outside the W3C are permitted if adequately defined" without defining what 'adequate definition' comprises.
In addition to having appropriate semantics, PDFs should have appropriate text equivalents for non-text content.
The PDF should pass Adobe's accessibility check (I am not aware of any third-party PDF validators).
The Errata state "Do not provide a sitemap or table of contents unless the site cannot be understood or navigated without it" without any justification. There is no reason to prohibit the provision of a sitemap, and they can often be useful without being essential. The checkpoint should be changed to "Provide a sitemap or table of contents if the site cannot be understood or navigated without it".
The removal of the requirement to group links and provide a means to bypass them is welcome. However, I do not interpret this as meaning that we are prohibited from using skip links where they are useful.
Although user agents can provide facilities to skip to certain element types such as headings, they cannot know where the content starts because there is no means by which do define that semantically using HTML. Until there is such a means, it is entirely reasonable for authors to provide 'skip to content' links.
Site Map
Copyright © 2002-2005 Test Partners Ltd. All rights reserved.
First Accessibility - accessibility services from Test Partners Ltd