Talk:Verdier duality


 * The section relating Verdier and Poincare duality is totally incomprehensible. Generally, this article suffers from the notation being too terse.--345Kai (talk) 23:41, 29 August 2011 (UTC)


 * Rewrote the section relating Verdier and Poincare duality. Tried to improve the flow of thought overall and deleted the section on notation since some hypotheses were introduced and never used. Added Kashiwara and Schapira as reference. Upgraded the class to a B. Jmcurry (talk) 00:47, 19 April 2012 (UTC)

Dualizing Complexes
There should be a separate page for discussing dualizing complexes in the context of algebraic geometry. A good *new* reference for the subject is contained in https://www.math.bgu.ac.il/~amyekut/teaching/2016-17/der-cats-IV/public63.pdf He constructs the dualizing complex for commutative rings on the bottom of page 218 and the top of 219.

correction for intertwining of functors
It's very easy to derive the property as stated now $$D(Rf_{!}(\mathcal{F})) \cong Rf_\ast D(\mathcal{F})$$. I therefore *assume* that the version stated before (with ! and * intertwined) was simply a typo, but I'm quite ignorant about these things. Any objections to the change? Amitushtush (talk) 23:36, 21 February 2018 (UTC)

X finite dimensional?
I think "If X is a finite-dimensional locally compact space..." needs to be clarified. What does finite dimensionality mean? I assume some sort of cohomological finiteness...? Amitushtush (talk) 23:55, 21 February 2018 (UTC)

According to Sheaves On Manifolds (Kashiwara & Schapira, 1990) Definition 3.1.16(i) & (ii) on page 148 the assumptions are that the space is locally compact and Hausdorff, $$f_!$$ has finite cohomological dimension, and that the space X has finite c-soft dimension (page 133 exercise 9(b); partially defined by exercise 9(a)). A c-soft sheaf (defined on page 104) is a compact analogue of flabby (or sometimes flasque) sheaf.

use of inaccessible cardinals??
Gobal Verdier duality involves a comparison of homsets in a derived category which in general might not be actual "sets" in the conventional sense as derived categories are usually by construction "large" categories (relative to a fixed universe). i.e. the derived category in general isn't locally small. If it is necessary it might be good to add a note that this result could rely on more than ZF by requiring the existence of at least one (strongly) inaccessible cardinal. I believe it is referred to by some as the "Grothendieck axiom".