How to Use
Where to enter
The path is linear by construction, but nobody arrives at zero. Use the table to find an honest entry point.
| If you… | Start at |
|---|---|
| Ship features and want to understand structural decisions | Level 01, from the beginning |
| Structure code well but have never designed a whole system | Level 01 skimmed, then Level 02 |
| Have designed systems, but only monolithic ones | Level 03, returning to 01 when vocabulary is missing |
| Work with distributed systems and want to close gaps | Level 04, then the Level 05 disciplines you lack |
| Are preparing for system design interviews | Level 03, then section 22, then the case studies |
| Are an architect wanting to work above the system | Levels 06 and 07, using the rest as lookup |
Skipping levels is legitimate. What does not work is skipping and never returning when a concept shows up as an assumption — every document declares its prerequisites precisely to make that visible.
Pace
There is no deadline. There is a trap: reading too fast produces recognition without capability. You read about eventual consistency, you agree, and six months later you cannot decide whether your case tolerates it.
The signal that the pace is right is being able to apply the concept to a system you know. If you cannot, reread — or go to the exercise, which exists precisely to force that.
The exercises
Exercises are not reading checks. They are where the reasoning happens.
Each one carries a collapsed Discussion section. It is not an answer key: it is one possible line of reasoning, with its assumptions stated. Opening it before trying wastes the exercise, because the value lies in discovering which constraint you had not considered — and that only surfaces after you have considered some.
Exercises 03 through 07 share one system and evolve it. Do them in sequence: the central lesson is feeling the cost of decisions taken early.
The case studies
Same rule, stronger. Read context, requirements and constraints. Stop before the architecture options. Sketch your own in twenty minutes. Only then continue.
The goal is not to be right. It is to compare your framing of the constraints with the one in the text.
About the trade-offs
Section 20 is the conceptual backbone of the material. If you only have time for one section, that is the one — but it pays much more after Level 04, because half of the trade-offs listed there only exist in distributed systems.
What this material does not do
It does not hand you answers. Most documents end with a decision conditioned on constraints, not with a recommendation.
That is deliberate. The capability this path develops is precisely stating those conditions — and material that delivers conclusions without them trains the opposite habit.
Language
The canonical content is in Portuguese. The English version is translated progressively; pages not yet translated fall back to Portuguese with a notice. The per-document state lives in the roadmap.