I had a brief look at "On Defining Proofs of Knowledge" by Bellare and Goldreich and I am a little confused by their definitions.
I was under the impression a negligible function $f$ was defined as something like $$\forall\ polynomials\ p\ \exists k\ s.t.\ \forall x > k: f(x) < \frac{1}{p(x)}$$
And that non-negligible meant simply that it was not negligible. The paper however states: "Put in other words negligible is not the negation of non-negligible!" (p. 5) And this seems to be based on the definition "a non-negligible function in $n$ is a function which is asymptotically bounded from below by a function of the form $n^{-c}$ for some constant $c$" (p. 4) which I am losely translating as $$\exists\ polynomial\ p\ and\ k\ s.t.\ \forall x > k: f(x) > \frac{1}{p(x)}$$ With the difference being functions which are somehow alternating.
This is a bit of a weird question because the mathematics seem to be clear but I am confused by what the common usage is. And is this something that is generally important? I've never seen it discussed elsewhere.