> You are missing the fact that it's impossible to make all functions
> related to regexps begin with "regexp-", because some of them are also
> related to lists or strings or buffers or email or projects. Or are
> you suggesting to have dozens of different aliases for each function,
> one each for every class to which it could possibly belong?
>
> How big is this "some of them" you are talking about? Are they really about regexp, or another topic?
"C-u C-h a regexp RET" and see for yourself.
I see very few candidates where it's unclear... and almost all the results is already properly namespaced (under dired when it concern dired, etc).
Remember the proposal is to unify what is the building blocks of regexps under the regexp- namespace. NOT to put everything that contains "regexp" under that namespace. That means how to create regexps, how to apply them to strings, how to get the match results. There might be things where it's hard to decide which is the main topic but that is more related to generic functions, which can remain generic for now until we figure it out.
The proposal is to start where it's obvious, not to sort it all out for all cases.
> > The above are _commands_ not functions. The equivalent would be
> > "C-h a regexp search RET", which does do what you want, I think.
> >
> > This again gives me a giant list where a majority of what is listed is not what I'd like.
>
> "Giant"? I see 11 hits (in "emacs -Q").
>
> In "Emacs -Q" you are correct. Without "-Q" the list is here https://www.ideone.com/35kbNi. On that list only
> line 116-127 are relevant to what I search.
You didn't type "C-h a", because that only shows commands.
Okay my bad, on my setup that binding runs `counsel-apropos` and not `apropos-command`. This is why you thought the list was reasonable and I thought it wasn't. Sorry.
Philippe