Catalog
A set, and one thing out of it. A page here either shows the set narrowed and paged, or the one entry a reader found in it.
The front door#
The front door to a large set, and not a product pitch. A reader who arrives here looks for one thing out of hundreds. So the first control is the search, and the first thing under it is the set itself. What a page like this owes is counts. An adjective is something anybody can write and a reader cannot check; a figure they can.
The catalog#
Several hundred small, uniform things a reader arrives at with half the name. A reader finds them by shape, so the wall is tiles and not cards. The drawing is the box, and the identifier under it holds back until the shape turns up. Search is the control at the centre of the page, and the set comes in pages. The facets narrow the wall for real, and which items show is the page's state, never the wall's.
One entry#
One entry of a catalog, in full. What the reader came for is the identifier
and the confidence that this is the right one. So the name stands as the
machine text it is, and the item shows in both modes rather than on two
colours somebody typed in. Every ground here is the system's own, forced
onto a subtree. A page that hardcodes #ffffff has left the theme.
The list#
News, releases, references and search results are the same page with different rows. It proves the set, not one entry, and the state a list page usually skips: a filter that matches nothing, with how much it read.
The results#
The query stays in the field, so a refinement is not a retype. The facets say how many stand behind each, so a narrower search is a decision. A source that answered with nothing says so.
An answer#
An answer always carries its source, its version binding, and what it leaves out. Every state in Guidelines → States carries exactly that.