W3C home > Mailing lists > Public > w3c-wai-ig@w3.org > January to March 2017

RE: Heading structure with 1.3.1 Info and Relationships and 2.4.1Bypass Block

From: <alands289@gmail.com>
Date: Tue, 3 Jan 2017 15:05:21 -0500
Message-ID: <586c0400.4269370a.ac818.26f1@mx.google.com>
To: "Denis Boudreau (gmail)" <dboudreau01@gmail.com>
Cc: Glenda Sims <glenda.sims@deque.com>, "McSorley, Jan" <jan.mcsorley@pearson.com>, WAI Interest Group <w3c-wai-ig@w3.org>

Thank you for your valuable feedback and information.

I can never stop learning.


Alan Smith

From: Denis Boudreau (gmail)
Sent: Tuesday, January 3, 2017 2:27 PM
To: alands289@gmail.com
Cc: Glenda Sims; McSorley, Jan; WAI Interest Group
Subject: Re: Heading structure with 1.3.1 Info and Relationships and 2.4.1Bypass Block

Hi Alan,

Example no2 on H42 has the following heading structure, where there’s a <h2> in the content BEFORE the <h1> is rendered, which implicitly indicates that using a h2 before a h1 (therefore not following the expected hierarchical structure of headings) is acceptable.
Ref.: https://www.w3.org/TR/WCAG20-TECHS/H42.html#H42-ex2:

	<h2>Site Navigation</h2>
	<h1>Stock Market up today</h1>
	<h2>Related links</h2>

With regards to your comment on F43, I think what F43 advises against is heading outlines that no longer reflect the actual structure of the content… like when heading elements are used to make content big and bold, but that content is not, in fact “heading material”. As long as the headings used for markup really represent content that is meant by the author to be headings, F43 would not apply.

So what that leads me to understand is that while it’s ideal to never skip a heading level and to hierarchically nest your headings so they are 100% accurately defining the document outline, it is acceptable from a purely WCAG compliance perspective to have a somewhat "wonky" heading structure, as long as the headings used really mark up heading content. 

That being said, I take back what I said about the WebAIM survey. From memory, I thought most users didn’t mind much about the heading levels, but I went back to double check, and it turns out they d care… so maybe IT WOULD be a good idea to consider that as part of WCAG 2.1 and/or Silver: http://webaim.org/projects/screenreadersurvey4/#levels. Here’s what they asked in Survey no4:

When navigating a web page by headings, how useful are the heading levels (e.g., "Heading 1", "Heading 2", etc.) to you?
• Very useful: 47.4%
• Somewhat useful: 34.7%
• Not very useful: 12.2%
• Not at all useful: 3.5%
• I don't know: 2.2%
But I believe that it still doesn’t change the fact that there isn’t anything in SC 1.3.1 that allows us to fail a page for compliance over skipped heading levels. I would still highly recommend to my clients that they fix this whenever they can, I would teach my students to align to a hierarchical structure as much as possible and advise that they bend over backwards to do so (especially now that I stand corrected with regards to WebAIM’s survey) , but whenever it comes down to bare bones compliance, I have no choice but to acknowledge that it is just a best practice and that this is in no way “mandatory” based on the normative documentation of WCAG 2.0.


Denis Boudreau
» dboudreau01@gmail.com
» 514-730-9168

On Jan 3, 2017, at 13:55, <alands289@gmail.com> <alands289@gmail.com> wrote:

Please confirm, as I don’t see any hierarchical break in the example in H42. 
>From the text of H42:
In the following example, headings are used in a hierarchical layout with h3 as a subsection of h2, which is a subsection of h1.
<h1>Plant Foods that Humans Eat</h1>
<p>There are an abundant number of plants that humans eat...</p>
<p> A fruit is a structure of a plant that contains its
<p>The apple is the pomaceous fruit of the apple tree...</p>
<p>The orange is a hybrid of ancient cultivated origin...</p>
<p>Banana is the common name for herbaceous plants ...</p>
<p>A vegetable is an edible plant or part of a plant other than a
  sweet fruit ...</p>
<p>Broccoli is a plant of the mustard/cabbage family ... </p>
<h3>Brussels sprouts</h3>
<p>The Brussels sprout of the Brassicaceae family, is a Cultivar
  group of wild cabbage ...</p>
<h3>Green beans</h3>
<p>Green beans have been bred for the fleshiness, flavor, or
  sweetness of their pods...</p>
Skipped a heading level DOES fail 1.3.1 based on F43: https://www.w3.org/TR/WCAG20-TECHS/F43.html
F43: Failure of Success Criterion 1.3.1 due to using structural markup in a way that does not represent relationships in the content
The objective of this technique is to describe a failure that occurs when structural markup is used to achieve a presentational effect, but indicates relationships that do not exist in the content. This is disorienting to users who are depending on those relationships to navigate the content or to understand the relationship of one piece of the content to another
Alan Smith
From: Denis Boudreau
Sent: Tuesday, January 3, 2017 1:23 PM
To: Glenda Sims
Cc: McSorley, Jan; WAI Interest Group
Subject: Re: Heading structure with 1.3.1 Info and Relationships and 2.4.1 Bypass Block
+1 to what Glenda and Phil said. Not skipping headings is a great best practice, but that’s all it is… a best practice. There’S nothing in SC 1.3.1 that allows you to fail the Success Criterion based on skipped heading levels. The WCAG WG even uses examples where they’re breaking the hierarchical structure of their headings in the techniques (see H42): https://www.w3.org/TR/WCAG20-TECHS/H42.html.
I’m not so sure that making it part of 2.1 or Silver would be such a great idea, based on the feedback we get from users and the WebAIM surveys though. It doesn’t seem like it’s such a big deal for most screen reader users (those who are most likely to be impacted by this). I’d love to see more user-supported data on this.
On Jan 3, 2017, at 12:40, Glenda Sims <glenda.sims@deque.com> wrote:
Jan, Phill, et al
I agree with Phill, in that skipping heading levels does not fail 1.3.1.  I think it would be a wonderful addition to WCAG 2.1 or Silver.  
WCAG 2.0 is a minimum bar...it is not all best practices.  There were reasons that skipped heading levels were not included in WCAG 2.0 SC 1.3.1. 

glenda sims    |   team a11y lead   |    deque.com    |    512.963.3773      

web for everyone. web on everything. -  w3 goals
On Tue, Jan 3, 2017 at 11:18 AM, McSorley, Jan <jan.mcsorley@pearson.com> wrote:
With all due respect, I don't agree that skipping headings is "an equal usability issue for everyone." I believe that it is much more time-consuming, and potentially confusing, for a person without sight to determine what is happening with heading structures that skip heading levels than it is for a person with sight.

I believe that heading structure for people who rely on the use of screen reading technology means something more than people who are able to see visual headings.  People who can see the headings, don't rely on them for navigation in the same way people without sight do.  I think that it is unfortunate that consensus could not be met to use best practice with heading structure.  It is certainly a fairness issue in assessment and I believe that it has a much bigger impact on usability than people think it does.

Jan McSorley
VP, Accessibility
Psychometrics and Testing Services

400 Center Ridge Drive, Suite E
Austin, TX  78753
M - (512) 673-9569
Twitter: @Jan_McSorley
Skype:  jan.mcsorley
Learn more at pearson.com

We put a man on the moon in the 1960's ... surely we can make information technology fully accessible to people with disabilities.  It can be done ... it must be done ... it will be done!
On Tue, Jan 3, 2017 at 11:01 AM, Phill Jenkins <pjenkins@us.ibm.com> wrote:
1.3.1Info and Relationships: Information, structure, and relationshipsconveyed through presentationcan be programmatically determined or are available in text. (Level A) 

1. I do not believe skipping a heading level is an issue for assistive software users any more than it is for a non-assistive software user.  In other words, it is an equal usability issue for everyone, but not a 1.3.1 success failure. 

2. I agree that headings (H1-H6) should be used structurally as the HTML markup language semantics intended them to be used, but 1.3.1 does not require the use of headings, does not required that headings be used all the time in all content, nor that when heading are used that they be used sequentially, nor that there only be one and only one heading level 1 per page.

3. 1.3.1 doesn't require that presentation include structure, but that if and when it does present structure and relationships through presentation, that the same information is available through mark-up (or available in text). 

If we want to "add" a new success criteria that requires that if and when heading levels are used, that they be used sequentially, but we have to allow for many situations that are consistent with the intent of the HTML specification.  For example:
a. Page that is quoting a part of a page that doesn't include a heading level 1 should be allowed.  In other words starting with a heading level 2 or 3.

b. Heading level 2 can follow a heading 3 or 4 if it is going back up a level.  For example:

c.  Page that doesn't include any heading level 3's, but uses a consistent heading level 4 for all footers across a set of pages should be allowed.  In other words, it is OK to skip a heading level on a page in some situations.  

So, with these examples we can begin to see why the working group could NOT reach consensus to ALWAYS require the use of headings, ALWAYS starting heading level 1 and NEVER skipping a heading level. 
Phill Jenkins
Senior Engineer & Accessibility Executive
IBM Research Accessibility
Received on Tuesday, 3 January 2017 20:05:55 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 3 January 2017 20:05:55 UTC