W3C home > Mailing lists > Public > w3c-wai-ig@w3.org > January to March 2018

Re: Having h5 after h2 is a violation as per 1.3.1 info & relationships. - was: WCAG violations or accessibility enhancements

From: Olaf Drümmer <olaflist@callassoftware.com>
Date: Sat, 3 Mar 2018 11:57:45 +0100
Cc: Olaf Drümmer <olaflist@callassoftware.com>, Harry Loots <harry.loots@ieee.org>, Jonathan Avila <jon.avila@levelaccess.com>, "Patrick H. Lauke" <redux@splintered.co.uk>
Message-Id: <99C6F0BD-65D6-4A81-80C3-0260D77DE434@callassoftware.com>
To: w3c WAI List <w3c-wai-ig@w3.org>
>> While i agree that skipping levels is not a WCAG violation,  it should be emphasised that this will cause a usability problem for screenreader users. 

It is only coming across as a usability issue because screen reader implementations chose a certain approach for navigation through a page or document using 'headings' as landmarks.

As it would not be too difficult for a screen reader implementation to figure out that a 'heading level has been skipped' (e.g. an h3 followed by h5): 
- it could still offer navigation on the next lower available level (in this case h5, not h4) and make it more probable that a user relying on this mode of page/document navigation does not miss out on a part of the page/document because of that skipped heading level 
- a screen reader implementation could also indicate in some fashion that in such a situation a heading level 'is missing'

It is a fairy tale that sighted users are seeing a clear structure of heading levels. All sighted users usually see is relative prominence of headings (exception: headings are numbered according to their level). What I am describing above is based on a concept of 'relative heading level'. Following my proposal would provide equal / same quality access to screen reader users as it exists for sighted users.

Furthermore, it is easier to adjust a few screen reader implementations  (and get the job done forever) than to convince the world out there that an h3 followed by an h5 is a bad thing (and to be honest: there'd be many more urgent issues to address than this one...). 

My 2 cents.

Olaf

PS: ...and as an aside: 
PDF/UA-1 (the ISO standard for accessible tagged PDFs) decided to prohibit skipped heading levels. For a number of reasons this provision will most likely be removed from PDF/UA-2 (the next version of PDF/UA, currently being worked on in ISO). One of the reasons is that there are documents out in the world that intentionally and rightfully skip heading levels. For example, in some government agencies in Germany, certain types of long documents have a clearly designed heading level structure where in some cases a heading level has to be skipped (because at that level does not apply). As the rules for these documents are very old already, and as nobody ever hseems to have had a real problem with this, it is an undue burden for them to adjust their approach. As they wish to be conforming with PDF/UA-1 they are currently inserting a fake heading that essentially says 'heading intentionally left empty'. Now, once I see a hack like this, all the alarm bells ring to tell me there probably is a conceptual issue... (in this case in PDF/UA-1, with its prohibition of skipped heading levels)


> On 3. Mar 2018, at 10:10, Harry Loots <harry.loots@ieee.org <mailto:harry.loots@ieee.org>> wrote:
> 
> While i agree that skipping levels is not a WCAG violation,  it should be emphasised that this will cause a usability problem for screenreader users. 
>  E.g. if a screenreader user cycles through the headings (h1, h2's then h3's, etc. ) and the screenreader reported nothing for a next level of heading (i.e.,  h3 and h4 skipped), then the user will most likely stop cycling through headings,  having assumed there were no further headings. 
> In this case the screenreader user will not be presented with an equivalent 'picture' of the page as we,  the sighted user will perceive it. 
> If it's not an explicit rule of WCAG, then it's high time we made it a rule! 
>  
> Kindest regards 
> Harry
> 
> 
> On 3 Mar 2018 03:42, "Jonathan Avila" <jon.avila@levelaccess.com <mailto:jon.avila@levelaccess.com>> wrote:
> > We could call it an "outline algorithm" perhaps. And then wait for browsers to actually implement it in a meaningful way too?
> 
> Yes, too bad this wasn't sent on February 2nd -- perhaps the second time around is a charm and we'll also be better piano players.
> 
> Jonathan
> 
> -----Original Message-----
> From: Patrick H. Lauke [mailto:redux@splintered.co.uk <mailto:redux@splintered.co.uk>]
> Sent: Friday, March 2, 2018 9:59 AM
> To: w3c-wai-ig@w3.org <mailto:w3c-wai-ig@w3.org>
> Subject: Re: Having h5 after h2 is a violation as per 1.3.1 info & relationships. - was: WCAG violations or accessibility enhancements
> 
> On 02/03/2018 14:51, Glenda Sims wrote:
> > Crazy idea o' the week.  This problem of heading structure never
> > happens to us in ordered lists..because we let the browsers render the
> > list numbers.  Wouldn't it be super cool if we could have a similar
> > construct in html for headings.
> >
> > I've only had one cup o' coffee....so this may not be the best idea I
> > ever had...
> >
> > Example of dynamic heading structure
> > <oh>  Glenda's make believe ordered headings
> >     <hi>  Glenda's make believe heading item </hi>
> >     <oh>
> >          <hi> Another make believe heading item </hi> <!--#include
> > file="includewithheading.html" -->
> >          <hi>The last heading in this example</hi>
> > <!-- end of include -->
> >      </oh>
> > </oh>
> 
> We could call it an "outline algorithm" perhaps. And then wait for browsers to actually implement it in a meaningful way too?
> 
> P
> --
> Patrick H. Lauke
> 
> www.splintered.co.uk <http://www.splintered.co.uk/> | https://github.com/patrickhlauke <https://github.com/patrickhlauke> http://flickr.com/photos/redux/ <http://flickr.com/photos/redux/> | http://redux.deviantart.com <http://redux.deviantart.com/>
> twitter: @patrick_h_lauke | skype: patrick_h_lauke
> 
Received on Saturday, 3 March 2018 10:58:17 UTC

This archive was generated by hypermail 2.3.1 : Saturday, 3 March 2018 10:58:18 UTC