I found a note about compactly generated. This is the article http://www.math.uiuc.edu/~franklan/Math535_1205.pdf.
I worry whether the proof of Proposition 2.4 is true. I not understand why the function $\tilde{q}$ from $X$ to $kY$ is continuous. I thought that the Proposition 2.4 is valid whenever $X$ is Hausdorff. But in that note, $X$ may be not a Haudorff space.
Can someone tell me?
That's what you could call the universal property of compactly generated spaces (I'll call them $c$-spaces for convenience):
Depending on how much category theory you know, you may understand this as the fact that the inclusion functor from the category of $c$-spaces to the category of all spaces has a right adjoint, the $k$-ification, with the counit being the identity $kY\to Y$.