RE: Skipped headings

Thank you all, apologies for having been quiet on this while travelling - There's lots of food for thought in your responses and I'll go away and consider them.

As I noted earlier - the checkpoint requires that the visual and semantic match - if the visual doesn't have a particular level then it's not, to my mind a failure of 1.3.1 asd long as there's not a confusing or incorrect relationship between elements. If two things are the same importance they have the same heading level - if the first was has a sub component that skips a level it is still showing a clear subordinate relationship to the one it follows.

Again thank you all

Kevin

From: Kevin Prince <kevin.prince@fostermoore.com>
Sent: Wednesday, April 10, 2024 1:38 PM
To: Steve Green <steve.green@testpartners.co.uk>; Adam Cooper <cooperad@bigpond.com>; w3c-wai-ig@w3.org
Subject: RE: Skipped headings

An argument I've often had with my screenreader colleague - the heading structure exactly reflects the content which doesn't have, or need an obvious H2 (usually). If the heading levels are the same visually and semantically, whether there is a skip or not, then surely it ceases to be an access problem and may, but usually isn't, be a UX issue.

Kevin

K

Kevin Prince
Product Accessibility & Usability Consultant

Foster Moore
A Teranet Company

E kevin.prince@fostermoore.com<mailto:kevin.prince@fostermoore.com>
Christchurch
fostermoore.com<http://www.fostermoore.com/>
From: Steve Green <steve.green@testpartners.co.uk<mailto:steve.green@testpartners.co.uk>>
Sent: Thursday, March 21, 2024 7:48 PM
To: Adam Cooper <cooperad@bigpond.com<mailto:cooperad@bigpond.com>>; Kevin Prince <kevin.prince@fostermoore.com<mailto:kevin.prince@fostermoore.com>>; w3c-wai-ig@w3.org<mailto:w3c-wai-ig@w3.org>
Subject: RE: Skipped headings

CAUTION: This email originated from outside of the organization.

It's actually worse than that. In most places, WCAG states that the heading levels should reflect the content structure, which leads to the dichotomy described in the initial post. However, in at least one place, it states that the heading levels should reflect the importance of the content, which is entirely different and far more arbitrary.

Steve Green
Managing Director
Test Partners Ltd


From: Adam Cooper <cooperad@bigpond.com<mailto:cooperad@bigpond.com>>
Sent: Thursday, March 21, 2024 5:46 AM
To: 'Kevin Prince' <kevin.prince@fostermoore.com<mailto:kevin.prince@fostermoore.com>>; w3c-wai-ig@w3.org<mailto:w3c-wai-ig@w3.org>
Subject: RE: Skipped headings

It's a good question, Kevin, but, that there is no definitive answer for it 20 years in to WCAG 2.x is not so ...

WCAG is silent on the contiguity of headings, whether nesting is conditional or typed, whether h1 must be the first in a view, whether there can be more than one h1 in the same context, whether font-size must match heading level , whether headings can be repeated in the same context, even tests for determining whether headings describe topic or purpose, and so on.

And then there's the dilemma as to whether techniques are merely 'guidance' or a backdoor to affirming a success criterion... we've all been there: "I'm sure that this is a failure ofXYZ - I'll use technique PQR for building my argument'.

That there is more than one way to implement a heading structure as you demonstrate means either these variations have no impact on people using the web or they do as your screen reading colleagues suggest and it is not addressed either by specification or by guidance.

As a screen reader user, I'd say that all ingredients should be the same heading level because it makes more sense to me that they are typed rather than nested contiguously or (so-called) logically.

But that's really just a preference because then I can skip through the ingredients for each meal to see if I have the items in the  fridge ...

And, if you imagine one of those typically imprecise usability testing questions like 'how  would you find ingredients for the chicken taco' my and everyone else's' preferences are lost, and all that remains is a comparison of how quickly this task is completed or otherwise.

To enumerate every use case and preference would be pretty tricky, I imagine, which is why WCAG 2.x concentrates on the 'effectiveness' component of ISO9421, but it would be great if WCAG 2.x could have evolved to include more if not more definitive guidance to reduce this uncertainty and better support accessibility practitioners working at the coalface (and this is not about more education and outreach).

I appreciate and accept that writing such guidance is no easy task, but the structure of WCAG 2.x (and the proposed structure of WCAG 3.0) with such a strong focus on evaluating rather than design and development may be the cause of this uncertainty - wouldn't it make more sense to concentrate on upskilling  those at the beginning of a lifecycle rather than arriving at or near the end to deliver a bunch of bad news merely to mitigate risk?






From: Kevin Prince <kevin.prince@fostermoore.com<mailto:kevin.prince@fostermoore.com>>
Sent: Thursday, March 21, 2024 1:47 PM
To: w3c-wai-ig@w3.org<mailto:w3c-wai-ig@w3.org>
Subject: Skipped headings


Disclosure - I'm sighted and look for logic not slavish adherence to a rule for the sake of it.

Firstly my colleagues who use screenreaders say that skipped headings do cause them to be confused: is that because  we, as an industry, have made the case too strongly that nesting must be seamless? Should we not be reflecting the relationships in the content not looking to plug apparent gaps.

Take this example (which is an extension of the Webaim example on their page)
H1: My Favorite Recipes
H2: Quick and Easy
H3: Spaghetti

              *   Ingredients of spaghetti

H3: Hamburgers
H3: Tacos
H4: Beef Tacos

              *   Ingredients of Beef tacos

H4: Chicken Tacos
H4: Fish Tacos
I would strongly argue that visually, logically and semantically both sets of ingredients should be the same heading level - they are the same thing. However, because there are sub-types of Taco the Taco ingredients drop to H5. As the semantics are supposed to convey relationships then surely spaghetti ingredients are also H5 irrespective of where the sit respective to Spaghetti?
I'm genuinely looking to understand why grouping similar things is wrong purely to maintain a lack of gaps. Visually you would clearly not miss the H4 under Spaghetti because it's not needed - why is it so important for a screenreader?

My reading also shows it's strong best practice but not, as far as I can tell, a failure of 1.3.1. And the essence of 1.3.1 is to convey the relationships shown visually (and visually all the ingredients are peers.

Sorry - really  would like to hear the consensus on how this should be marked up and why?

Kevin



Kevin Prince
Product Accessibility & Usability Consultant
[cid:image001.jpg@01DA8B4D.76BFB9B0]
E kevin.prince@fostermoore.com<mailto:kevin.prince@fostermoore.com>
Christchurch

fostermoore.com<http://www.fostermoore.com/>
This email and its contents are confidential. If you are not the intended recipient, you should contact the sender immediately, you must not use, copy or disclose any of the information in the email, and you must delete it from your system immediately.



Kevin Prince

Product Accessibility & Usability Consultant



E kevin.prince@fostermoore.com

Christchurch

fostermoore.com

This email and its contents are confidential. If you are not the intended recipient, you should contact the sender immediately, you must not use, copy or disclose any of the information in the email, and you must delete it from your system immediately.