WD08-1 suggests semantics measurable, but are they? #24
Description
In WD08-01 Validate Source Code for Semantic Accuracy, we suggest HTML validation checks syntax and semantic accuracy.
I'd say the first is correct, the HTML validator will check syntax and tell you if you miss required attributes or get testing wrong.
But the second, semantic accuracy, is not checked. For instance, if an author wants to display this list:
- Whisk eggs
- Fry mixture in a pan
- Add salt and pepper
they could mark up the list with <ol> (ordered) or with <ul> (unordered). It depends on whether they intend the order to matter. Are they okay with adding salt and pepper at the start or whisking at the end?
The same goes for a lot of HTML elements, the procedure to check is manual, by looking at the content's structure and verifying if the HTML conveys it.
If others feel the same, I see a few ways of potentially resolving this:
- Rewrite 3.8.1 to be about valid HTML. This might make it more specific to sustainability too, as (very) large validation issues could impact page performance (while semantics have more impact on SEO and accessibility)
- Remove 3.8.1 and its corresponding technique; I feel we could consider this, as a similar criterion (4.1.1 Parsing) was also recently removed from the last version of WCAG, as (paraphrasing) browser parsers can deal sufficiently well with syntax issues