for me at least, 90% of the time i already do know part of the name of the symbol im looking for, i just need to fill in some blanks in my memory. aside from search, the most useful organization for me would be grouping by:
- initializers
- static methods (since these are often initializer-like)
- instance properties
- subscripts
- instance methods
- static properties
- operators
in a well-designed API, the kind of a symbol (e.g., subscript) should already provide some hint as to its role.
if there too many instance methods to sort through, i am usually more interested in what protocol interface they are associated with than what “topic” they fall under. however, this type of organization is hard to implement with the information provided by the symbol graphs.
a site map is a search engine concept, like a robots.txt. unless you are referring to something different?
browsers cache favicons by default, even if no etag is provided. so you can serve a different favicon for each page if you want, but there’s no guarantee the user will see it.