Ten decisions a database engineer has to make. Each page picks a side, names the price, and says what breaks when the choice is wrong. One decision is written so far. The rest are planned, in the order I intend to write them.
Each decision page follows the same shape: what the choice is, why it matters, how each engine does it, when to use it, what breaks, the alternatives rejected, how to operate it, and the version notes. The numbers on a decision page are derived from stated assumptions until the Tested on box on that page says otherwise.