Definition 2.4.4 (Fiber and cofiber sequences in \(\An _*\)). A nullsequence in \(\An _*\) is a sequence of morphisms \[ X \xrightarrow {f} Y \xrightarrow {g} Z \] equipped with a specified identification \(g \circ f = 0\), called a nullhomotopy. Equivalently, it is a commutative square in \(\An _*\) of the form

Commutative square with maps X to Y labeled f, 0 to Z, X to 0, Y to Z labeled g.

Such a nullsequence is called:

  • a fiber sequence if this square is a pullback square in \(\An _*\) (or, equivalently, in \(\An \)). In this case we call \(X\) the fiber of \(g\) and write \(X \simeq \fib (g)\);
  • a cofiber sequence if the same square is a pushout square in \(\An _*\) (or, equivalently, in \(\An \)). In this case we call \(Z\) the cofiber of \(f\) and write \(Z \simeq \cofib (f)\).