I think the first (and probably second) assertion is incorrect.
You can only recognize some features "at a glance": if the structure is over a few screens away, you are not glancing. And in those cases, if you need to understand the structure at a glance you are better off with long element names that redundantly include the hierarchy information, such as <html.body.div.div.p> (which then may make the structures even less regular but more glanceable) or putting the hierarchy into some attribute (on containers).
So I think this is the 'black or white' logical fallacy. It is like saying, "eating dirt is better than eating arsnic" and ignoring that we can also eat Malaysian rendang. The conclusions only hold in a simplified world, not the real world. In this case, the black and white are that the only choices are glanceable and not glanceable, and declarative and non declarative (you should think of a positive word for non declarative: that there is none should be a red flag.) And this then leads to a second fallacy, of false equivalence, that declarative = glanceable.
Also, I think there is an improper shift ground too. Declarative is a term coming from computing as distinguished from 'procedural' or 'imperative'. An XML document is only declarative or imperative if it has those operational sematics applied, eg if it is an XSLT script. There is no sense in which a document is declarative absent knowledge of the element sematics. (And I dont see how the distinction between declarative and procedural knowledge in the domain of psychology applies here, except perhaps by tenuous analogy.)
I think the argument would be more sound by not mentioning "declarative" at all. If you mean regular and glanceable and comprehendable, use those terms.
But perhaps it is better to rope in studies of limits of human neuro-typical short-term or working memory, keying, and cognitive load, because it strikes me that while these syllogisms just attempt to capture and formulate craft knowledge, they are more properly assertions about human psychology?
{ASIDE} Indeed, it strikes me that while the statements appear to be saying that one kind of markup is better than another, it is also saying thst one kind of perception and comprehension is better than another, isnt it? That the thinking of a person who only knows what is in front of their eyes is better than a person who understands linearly (i.e preceding::*) or hierarchically (i.e. ancestor::*). Now I appreciate this, being exactly that kind who ptimarily knows only what is in front of my eyes, but it is not the only mode of legitimate effective cognition, is it?
Regards
Rick
