Sufficient Technique C27 not correct for SC 1.3.2

[markmagennis]

In Sufficient Technique C27 ‘Making the DOM order match the visual order’ it says “This technique is Sufficient to meet 1.3.2: Meaningful Sequence”. This is not always correct. Applying C27 may sometimes cause a failure of 1.3.2, not a pass. This would happen whenever a piece of content is visually positioned above another piece that is before it in the logical sequence. A typical and common scenario shown in the attached screenshot is where something is positioned above its heading.

The screenshot shows part of a web page that contains a number of eLearning resources. Some are videos, some are audiobooks. Each resource has a title, a type (“video” or “audiobook”), and a text description. The titles are big and bold and are headings in order to meet SC 1.3.1 Info and Relationships. So the meaningful sequence must have the type and the description coming after the title (the heading). If they were before the title they would be under the previous heading which would give them the incorrect meaning and fail SC 1.3.2. But visually the type (“video” or “audiobook”) is positioned above the title. In this case, following C27 ‘Making the DOM order match the visual order’ would mean placing the type before the title (its heading) in the DOM. This would cause a failure of 1.3.2, not a pass.

So I would suggest that it is wrong for C27 to state that it is “Sufficient to meet 1.3.2: Meaningful Sequence”.

[patrickhlauke]

as said on list:

I agree that C27 is taking a very generalised shortcut here for something that is quite nuanced, and will lead to possible false positives/negatives (depending which way you look at this). It should, at best, be heavily caveated ("in general", "in most cases", etc) and/or scoped to a very specific example only, with addition of a "...but it's not always the case".

[JAWS-test]

I wonder if the example in the screenshot would always be a violation of 1.3.2 (as long as the type is in the DOM before the heading)? Or would 1.3.2 be fulfilled if, for example, it is obvious from a grouping (e.g. article element) that the type belongs to the heading below it?

[detlevhfischer]

This is a common problem. it also appears in news sites where you have a visually prominent heading and text on top that gives it context, sometimes in a way that the big heading alone makes no (not enough) sense, so both are needed to get the meaning. We would usually recommend putting both top text and heading into one semantic heading container and style that visually to be on two lines. Others have argued for using the heading only on the visually prominent heading, and stick both top text and heading in to a header element.

If it makes auditory/programmatic sense to swap around top text and heading in the DOM, I would certainly not see a 1.3.2 failure. It seems this is one situation where there are no ideal solutions - and once you nail down a particular pattern, editors will come up with content that shows it up...