Want to take part in these discussions? Sign in if you have an account, or apply for one below
Vanilla 1.1.10 is a product of Lussumo. More Information: Documentation, Community Support.
Is there a place in the general framework of the page separation axioms for the notion of TD-space (apparently sometimes called T12?
This looks like a job for Toby! :-)
Not as far as I know.
Not every separation axiom fits into that scheme; see http://ncatlab.org/nlab/show/separation%20axioms#other_axioms_7 and https://en.wikipedia.org/wiki/Separation_axiom#Other_separation_axioms for places to list exceptions (although TD is on neither list). By the way, I have heard of another separation axiom called ‘T1/2’, but I never tracked down the reference and nothing came of the discussion: https://en.wikipedia.org/wiki/Talk:Separation_axiom#T1.2F2_spaces.3F.
I don’t really have a good understanding of TD-spaces.
I’ve started reading the new book on locales by Picado and Pultr. They point out that the TD axiom also allows one to recover a space up to homeomorphism from its frame of open sets, though in a different way from the sober axiom. Rather than recovering points as completely prime filters, as we do for a sober space, we find them as prime filters F such that there exists a,b with a∈F, b∉F, and b<a with nothing in between. They also give an example of two spaces X and Y with isomorphic frames of opens, such that X is TD but not sober, while Y is sober but not TD. (X is the cofinite topology on ℕ; Y is its one-point compactification.)
They also make the point, which seems reasonable to me, that sobriety (or, at least, the half of sobriety that isn’t T0) is less of a “separation” property and more of a “completeness” property.
Yes, I’m also reading that book. I follow their discussion of TD, but I still don’t feel that I really understand the concept.
I also agree that sobriety isn’t really about separation, but then neither is perfect normality. Both should still be on a list of ‘other’ separation axioms, since other people list them sometimes.
1 to 5 of 5