General Comments/Questions and Comments/Questions on WCAG Tests 70, 149, 150 - Tim Boland Feb 11 2005

NOTE:Using metrics mentioned at Jan 26 2005 WCAG TTF teleconference.

DISCLAIMER: My personal opinions only. Apologies in advance for any misunderstandings or misinterpretations on my part.

QUESTIONS on CONFORMANCE TEST PROCESS DOCUMENT: Should there be an email address given where those outside WCAG can submit candidate tests, should there be an initial screening of newly-submitted tests to determine basic appropriateness to the relevant WCAG success criteria, should just-submitted tests have status designation of "just submitted" rather than "unconfirmed", and should ?

GENERAL COMMENTS/QUESTIONS: Maybe each test should have in its content a self-explanatory test purpose /objective, and a direct link to the actual relevant success criteria of the WCAG2.0 specification within the content of the test. For example, for Test 70, the stated (purpose?-my word) "List items should not be used to format text" is in the title of Test 70, but is not explicitly identified as a test purpose, and there is no explicit link to the Guideline 3.1 Level 2 WCAG2.0 success criteria in the contents of this test (just to the Guidelines statement in general). The same is true for the Test 149 and Test 150.

Also, should there be more documentation/metadata associated with the test case itself? A more meaningful naming of test cases for easy identification (other than "Test70", Test149", etc.).


Test 70 - List items should not be used to format text

This test deals with removal of the MARQUEE element (an HTML extension for one user-agent category? - not supported by all user agents?). Therefore, can this test be reasonably be applied to all web resources or require uniform behavior regardless of disability ("success criteria" level categorization of "Conformance" section of WCAG2.0)? What about non-HTML web resources? XML? What should the behavior be for user agents that may not recognize the MARQUEE element? Maybe this test should be level 1 or optional if used at all? Even though the test deals with removal of the MARQUEE element, should initial support for MARQUEE be required for the test to run correctly. since the MARQUEE element would have to be recognized correctly?

My understanding is that the MARQUEE element defines an area in which visual scrolling will be used to display the content of the element, and that a MARQUEE may define text which moves from side to side (or "flying text"?). Thus, I do not understand how this test could be used to satisfy any of the three level 2 success criteria for Guideline 3.1. The first two of these success criteria deal with meaning of words in content, whereas MARQUEE is a tag name and appears to me to have relatively little to do with meaning of words, just with movement of text. The third success criterion deals with foreign language identification, which is not explicitly mentioned in the test as well. Thus, the wording (vocabulary) in the test does not appear to me to synchronize well with the previously-mentioned success criteria. Perhaps some additional explanation (or links?) should be given in the text as to any connection between movement of text and meaning of words? If I don't understand, perhaps others won't as well..

The test mentions "document" not "delivery unit" as is discussed in WCAG2.0. What is the definition of a "document" in this context? Perhaps there should be a link to such a definition. Perhaps there should also be a link to definition of "MARQUEE" element (if appropriate?), so that accessors would know how to get more information to be able to complete the test?

How does movement of text within a resource consistent with testing that list items should not be used to format text (is a "list item marker" meant in this context, or just the contents of a list element)? Even though "ol" and "li" can be children of a "MARQUEE" element, I still don't understand the connection between lists and MARQUEE in a general sense.

Is a prerequisite for Test70 the use of valid HTML according to the HTML4.0 specification? Is HTML being tested, or is accessibility being tested?

Are there any alternatives to MARQUEE for providing the same functionality? Perhaps CSS Positioning re: DHTML? Are there any alternatives that are supported by multiple user-agents or are sufficiently general?

MARQUEE only appears to apply to visual rendering. What about audio, etc.? So, is this test sufficiently general, or should it be marked as "visual-only"?

I think that at the end of Test70 "any accessibility check.." is too vague - don't you mean "any other WCAG test may be performed after this one", or do we want to offer additional more specific guidance? Are there any dependencies on other tests for this test?

I think that in Test70 "..related to technique.." is too vague. Perhaps "This test may be run using this technique" or "this test supports the use of this technique" might be better?

I don't see how the technique given "abuse of list elements" relates to the use of the MARQUEE element; this technique as given in Section 6.2 in the HTML Techniques document maps back to G4.1, not G3.1 as given in Test70, and deals with not using list elements for presentational effects (for example, do not use "ul" and "ol" for indentation effects). How does this relate to "MARQUEE"? BTW, Section 12.5 of the CSS2.1 CR may have some useful information about lists, and Section 12.4 fo the CSS2.1 CR describes automatic counters and numbering, which may have relevance also.


Test 149 - Nesting ordered lists should use STYLE to convey list depth

NOTE: Some of the previous comments/questions apply in this context as well.

Regarding title of this test, I thought that list depth of nesting ordered list information should be preserved in the delivery unit structure, and not be indicated using any presentational aids. However, this test title seems to me to indicate the use of style to convey list depth, which seems to me to be a violation of G1.3. What am I missing? I would think that given an ordered list structure, one could use CSS, XSL-FO (or another presentation mechanism?) to render in different ways, but the ordering and depth are separate from presentation, and still may be accessed even when different presentations are not used.

Does the word STYLE (in caps) in the title of the test mean the HTML STYLE attribute, or the STYLE element using CSS, or non-CSS presentation? Unclear to me..

Since the "Test Process" is to be filled in for this test, it may be difficult to determine whether this test satisifes the level 1 success criteria categorization given in "Conformance" of the WCAG2.0 WD. However, from the title alone, it may appear to me that this subject can be reasonably applied to all Web resources (delivery units?), and may deal with minimum accessibility as required.

I don't understand how, from the title of this test, use of this test would satisfy the second and third Level 1 success criteria of Guideline 1.3. The second success criterion deals with emphasis, and the third deals with color. Even the first success criteria, determining the structure, seems to me inconsistent with the use of presentation to represent such information. Again, perhaps some more explanation is needed in the test (or additional links?), or I am missing something?

I can see how this test (when it is filled out) may be consistent with the use of the ordered list technique given (from Section 6.1 of HTML Techniques Document), in that this technique may be employed in support of running the test, to satisfy the stated success criteria. Also, the mapping of Section 6.1 of HTML Techniques document back to 1.3L1 SC1 seems to me consistent with the reference to Guideline 1.3 in the test itself. The CSS2.1 Chapter 12 material mentioned previously may also be helpful in the context of this test.

Shouldn't this test reference a CSS technique from the WCAG2.0 CSS Techniques Document (rather than or in addition to an HTML Technique) if CSS is to be used? Again, what about use of CSS with XML to represent lists?

CSS2.1 allows a user agent to generate numbers automatically, rather than list the numbers explicitly in a numbered list. Content may be generated by elements with a value of 'list-item' for the 'display' property.

Automatic numbering in CSS2.1 is controlled with two properties : 'counter-increment' and 'counter-reset'. The counters defined by these properties are used with the counter() and counters() functions of the 'content' property.

There is no test process to review; do you want one suggested?


Test 150 - Use CSS, not images, to change list bullets

NOTE: Many of the comments for Test 149 also apply to Test 150. Since CSS is explicitly recommended for this test, some suggestions for using CSS to "change list bullets" and using other CSS list capabilities) follow.

What does it mean to "change list bullets" (different renderings of "list style markers")? Also, I think that this test should definitely link to a "related"? CSS technique in the CSS Techniques Document, since CSS is explicitly mentioned in the title for the test. I did access a general technique? titled "separate the structure of lists and list items from presentation" which seems reasonable. CSS lists functionality given following would seem to me to be consistent with the "ordered lists" Section 6.2 technique ("format ordered lists so their items can be followed logically") given in the HTML Techniques Document.

CSS2.1 offers basic visual formatting of lists. An element with 'display: list-item' generates a principal box for the elements content and an optional marker box as a visual indication that the element is a list item. The list properties allow style sheets to specify the marker type (image, glyph, or number), and the marker position with respect to the principal box (outside it or within it before content). They don't allow authors to specify distinct sylte for the list marker or adjust its position with respect to the principal box.

The "list-style-type" property specifies the appearance of the list item marker if 'list-style-image' has the value 'none' or if the image pointed to by the URI cannot be displayed. The 'list-style-image' property sets the image that will be used as the list item marker. When the image is available, it will replace the marker set with the 'list-style-type' marker. The 'list-style-position' property specifies the position of the marker box in the principal block box.

The 'list-style' property is a shorthand notation for setting the three properties 'list-style-type', 'list-style-image', and 'list-style-position' at the same place in the style sheet. Although authors may specify 'list-style' information directly on list item elements (e.g., "li" in HTML), they should be careful when doing so.

The test purpose for this test is to be done; do you want a test purpose suggested?