is `local x` essentially introducing or declaring?

[WalterMadelim]

There is a serious confusion associated with local x.
It's docstring reads vague.
or is actually the docstring is correct but the current julia's behavior is wrong? I don't know.
Please see this discourse thread https://discourse.julialang.org/t/about-the-definition-of-local-keyword/132903/10?u=waltermadelim

[KristofferC]

Could you summarize your confusion here so readers don't have to read through a long external forum thread.

[WalterMadelim]

function f(a)
   local a
end

function c()
   local a
   local a
end

Please explain the logics of these two programs, using the most detailed words.
Why the first one cannot be defined, while the second one can be defined and can be run without any error.

[WalterMadelim]

I think There Should Be Only One Entity Which Is Referred To By The Name a Within The Body Of c().

According to the current docstring of local a.

The intention of c is

function c()
    introducing_a_new_name `a`
    introducing_a_new_name `a`
end

So we end up having two "different" but actually the same names, both visible as a. This is weird.