This article is a topic stub; I will keep adding insights about dependency graphs over time.
There are complex situations involving subtle dependencies and cycles between different components.
These kinds of situations can be wickedly hard, especially in the absence of sensemaking tools that make these causal relationships legible.
For me, this is where dependency graphs come in. Sensemaking about the causal structure of the situation involves drawing a structure made up of nodes and directed arrows denoting a dependency. Once the graph has been laid out, there is an explicit order of what events to tackle, starting with the rootiest first.
Cycles are subtle and interesting. In a true chicken-egg cycle, one answer that sometimes cuts through the circular dependency is parallelism: do both at the same time. Sometimes working on both at the same time can made the needed impact to update both factors. Another option is searching for unrecognized dependencies upstream of both factors.
Dependency graphs are a powerful sensemaking tool, and I suspect they should be utilized for more use-cases.