> I think you are making a distinction between names that are
> core parts of the use of regexps, and names that contain 'regexp'
> because they stand for something that uses a regexp somehow.
>
> Is that right?
Yes.
> And you expect the former to have names that start with 'regexp'
> (although we have never had such a naming practice for data types).
>
> Is that right?
Yes! :-)
> Is the reason you expect the names to follow that pattern
> that you are coming from a language that uses abstract object tyoes
> where each type defines methods to operate on it? Do you wish you
> could ask, "Show me the operations defined on type 'regexp'"?
Yes, and also because in almost every other languages there are namespaces. Including other lisps (Scheme, Clojure). Even in Emacs Lisp the namespace concept is used, look at the all the `string-*` functions.
> But if it is just to make C-h f completion on 'alist' include 'assoc',
> we don't need real aliases. We could do it with aliaases that work
> only in completion of function names. We could make an alias
> ("alist" "assoc"), which would add "assoc" to the list
> of completions of "alist".
>
> These aliases would avoid the downsides completely. Would they help you?
Maybe, I see that as being a bit dangerous and unexpected for a lot of users tho. One expects plain text search in C-h f.