Why Computer Science Trees Grow Downwards
Description
The image is a snapshot from a computer science or mathematics textbook page. It displays a diagram labeled 'FIGURE 1.2. A decision tree for selecting a subset of {a,b,c}.' The diagram is a classic binary tree, starting with a root node at the top and branching downwards to leaf nodes, which represent all possible subsets of the set {a,b,c}. Below the diagram, there is a 'Remark' section with text that reads: 'A picture like this is called a tree. (This is not a formal definition; that will follow later.) If you want to know why the tree is growing upside down, ask the computer scientists who introduced this convention. (The conventional wisdom is that they never went out of the room, and so they never saw a real tree.)' The humor lies in this witty, self-deprecating explanation for a long-standing convention in computer science, playing on the stereotype of academics and programmers being disconnected from the natural world
Comments
7Comment deleted
We draw trees with the root at the top because traversing from the general case to the specific is just a recursive call. Also, it's easier to find the root of the problem when it's the first thing you see
Pro tip: any architecture diagram that grows roots upward is just foreshadowing - eventually the stack tips over and you’re pruning leaves at 2 a.m
After 20 years in tech, I finally understand why we call them 'decision trees' - because like real decisions in production, they branch everywhere, the root cause is always at the top, and by the time you reach the leaves, you're wondering how you got there and why everything is upside down
This textbook's savage burn perfectly captures why CS academics have such a fraught relationship with the physical world: they've been staring at upside-down trees for so long that when they finally venture outside, they probably try to traverse the park in O(log n) time and complain about the lack of proper parent pointers in the oak trees
Decision trees grow upside down because entropy splits downward - botany's loss, Gini's gain
CS trees grow upside down so the root cause is at the top - and so product doesn’t notice that “three toggles” just exploded into 2^n configurations
Peak CS: an upside‑down “decision tree” that’s really a power‑set generator (2^n leaves, zero impurity) - proof we’ll reinvent botany to avoid going outside