article_plan.txt

File: bottom-up/article_plan.txt

The formula for each article:

What I was trying to understand — the concept from the plan
What I actually did — the exercise, with the real commands and output
What broke or confused me — honest first-person moment
What it showed me — the insight, explained simply
The one-line takeaway


Example for Layer 2 (clone syscall):
Instead of explaining clone() theoretically, you write:

I compiled a 20-line C program that calls clone() directly. When it ran, the child process printed PID 1. Same machine, same kernel — but the process thought it was alone. That's the entire basis of container isolation. One syscall with flags.

Then show the C code, show the output, show the /proc/<pid>/ns/ comparison.

For LinkedIn specifically:
Each layer is too long for a post. Break it down further:

One post per key insight from the layer, not the whole layer
Layer 4 alone has 3 LinkedIn posts: OOMKill, CPU throttling, QoS classes
Hook line + 3-4 paragraphs + code snippet + takeaway

For the blog:
Full layer as one article — exercises included, real output, first person. Exactly the format we've been doing today with the PV/PVC articles.