Let $A$ and $B$ be complexes $R$-modules. Assume that $A^\bullet$ is bounded above and $B^\bullet$ is bounded below. Then there is a convergent spectral sequence $$\prod_{r\in \mathbb{Z}} \operatorname{Ext}^p_R (H^{r-q} (A), H^r (B)) \Longrightarrow \operatorname{Ext}_{\mathcal{D} (R\text{-Mod})}^{p+q} (A,B)$$ —see e.g. the very last result in Verdier's thesis.

If we are interested in complexes of abelian groups ($R = \mathbb{Z}$), then $\operatorname{Ext}^p_\mathbb{Z} = 0$, unless $p = 0,1$, and the above thing is a two-column spectral sequence, which gives us short exact sequences $$0 \to E_2^{1,n-1} \to H^n \to E_2^{0,n} \to 0$$ In particular, we have a short exact sequence $$\tag{*} 0 \to \prod_{i\in \mathbb{Z}} \operatorname{Ext}_\mathbb{Z} (H^i (A), H^{i-1} (B)) \to \operatorname{Hom}_{\mathcal{D} (\text{Ab})} (A,B) \to \prod_{i\in \mathbb{Z}} \operatorname{Hom}_\mathbb{Z} (H^i (A), H^i (B)) \to 0$$ The last morphism here is given by $f \mapsto (H^i (f))_{i\in \mathbb{Z}}$.

  1. This looks very natural and I wonder whether there is a direct way to get (*).

  2. Does (*) hold if the complexes are not bounded? I am interested in the case where they are both bounded below and not bounded above.

Thank you.


If $\text{Ext}^i_{\mathcal A}(-,-)\equiv 0$ for $i>1$ (that is, if ${\mathcal A}$ is a hereditary category) then any bounded complex is isomorphic to the direct sum of its cohomologies in ${\mathbf D}_{\mathcal A}$, i.e. $X\cong\bigoplus_n \Sigma^{-n}\text{H}^n(X)$ - see e.g. If the cohomology of two objects in the derived category are equal, are the objects isomorphic?.

Therefore, studying $\textbf{D}_{\mathcal A}(X,Y)$ as you do in $(\ast)$ naturally splits in two cases:

  1. $X=X^0,Y=Y^0\in{\mathcal A}$, i.e. both are stalk complexes in cohomological degree $0$.

    In this case, $(\ast)$ degenerates to $\textbf{D}_{\mathcal A}(X,Y)\cong{\mathcal A}(X,Y)$, i.e. the fact that ${\mathcal A}\to\textbf{D}_{\mathcal A}$ is fully faithful.

  2. $X=X^0, Y=\Sigma Y^{-1}$, i.e. $X$ and $Y$ are stalk complexes in cohomological degree $0$ and $-1$, respectively.

    In this case $(\ast)$ degenerates to $\textbf{D}_{\mathcal A}(X,Y)\cong\text{Ext}^1_{\mathcal A}(X,Y)$.

  • $\begingroup$ Thank you! Is there a way to do that for $\mathcal{D}^+$ instead of $\mathcal{D}^b$? $\endgroup$ – Lyosha Jun 12 '17 at 10:09
  • $\begingroup$ Indeed, your decomposition may be deduced without induction, hence the result doesn't need boundedness of complexes. So it solves my problem. $\endgroup$ – Lyosha Jun 12 '17 at 20:39

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.