# Homology of Chain Complexes from Free Resolution

Suppose I have an $R$-module $M$ and a free resolution $$\ldots \to F_2 \to F_1 \to M \to 0.$$ I apply an additive functor $f$ in $R$-$\mathbf{Mod}$ to the free resolution to get $$\ldots \to f(F_2) \to f(F_1) \to f(M) \to 0.$$ This is a chain complex, so I can take the homology of this. Will the homology be independent of the free resolution that I started with?

I believe the answer is yes. Here's my rough reasoning. First of all, the sequence $f(F_\bullet)$ is a chain complex because $f$ brings zero morphisms to zero morphisms by additivity. Any chain map between two free resolutions remains a chain map under $f$. Also, any chain homotopy between two chain maps remains a chain homotopy under $f$.

However, in all sources I have found, they would require $f$ to be right-exact, and the resulting homology after applying $f$ would be referred to as the left derived functor $L^i$. I am not sure if right-exactness is necessary in making the homology well-defined, or if it is needed to show that each $L^i$ is a functor, or if it is only needed to show that the sequence of $L^i$ forms a $\delta$-functor.

-
Exactness is needed for universality, I think. – Zhen Lin Jan 17 '14 at 13:50
You are correct that the homology is independent of the resolution. The main difference is that you will not have $H_0(F_\bullet) = f(M)$ unless $f$ is exact, but it will instead be something called the zero'th left derived functor of $f$. Some cases where this has shown up in real-world applications include where $f$ is completion with respect to an ideal. However, in some cases (e.g. where $f$ is left exact) you might not get an interesting answer. – Tyler Lawson Jan 17 '14 at 14:14

''Theorem $5.3$ shows that the right derived functors are of real interest only if $T$ is left exact. Indeed $R^n T$ may always be replaced by $R^n T'$ with $T' = RT$ which is left exact. Similarly the left derived functors are mainly interesting for functors which are right exact.''