[throwaway173738]

Just had this conversation with one of my juniors. He brought up scalability concerns with a PG queue and wanted to know if EG REDIS was a better choice. My whole tack was to lay out the actual context for the system—we’re going to soft launch, it’s for a hardware product, and we’ll have some one post-launch to make changes. Then I brought up my values since I’m basically in charge: maintainability, speed to first release, and ease of understanding. He made the decision that a PG queue was better for now because it’s easy to change our minds later, we already know PG, and we’re not anticipating fast scaling in this application.

I think a lot of the problem in industry is simply lack of rigor. We talk a big game about being software engineers but nobody takes the time to talk about values, requirements, trade-offs or business context for decisions. It’s a shame because these topics are the actual engineering of the system.