[structure] needs an overhaul

[zygoloid]

Looking through [structure], we find:

• [structure.elements] does not correctly describe any of the library clauses. There is no "Summary" subclause, no "Detailed specifications" subclause, the references to the C library are not collected together as is suggested therein. And what is an "element" anyway?
• [structure.summary]'s first sentence might be approximately correct, but the second sentence is not. And the second paragraph doesn't seem to provide any expositionary value whatsoever.
• [structure.specifications]: bullet 1.1 is not something we really do; paragraph 2 belongs on our wiki not in the working draft
• [structure.see.also] has a suspicious section name

There's lots of scope for improvements here.

[JohelEGP]

I had correlated the "Summary" and "Detailed specifications" with the [x.general] subclauses right after the [x] top-level Clauses and the descriptions of the library entities following the format in [structure.specifications], respectively. So I thought that [structure.elements] correctly described all the library clauses.

As for http://eel.is/c++draft/structure#summary-1.sentence-2, I thought those were the [x.syn] subclauses ("listing the headers specified in the subclause") and their content ("the library entities provided in each header.")

[structure.specifications]/1.1 is mostly done with classes and class templates (at least for the parts I usually refer to), but rarely with functions and function templates. For the latter, the only examples I can recall are in [utilities]: https://wg21.link/forward and https://wg21.link/declval, although they don't mention the name.

[jensmaurer]

Let's find out what we actually want to do here, and how much coordination with LWG we should endeavour.

[JohelEGP]

This might be a good opportunity to specify how to present concepts. The detailed specifications in [concepts] seem to follow a variation of those in [structure.specifications] better fit for concepts:

• Name and brief description
• Concept definition
• Semantics on expressions

[AlisdairM]

Should this be linked to #7399?
Should the milestone move to C++29, or do we still think that this is in scope for C++26?