I joined a financial firm that was building a trading system. This was about 15 years ago. The volume was not small. Their DB model was atrocious. I looked at

I joined a financial firm that was building a trading system. This was about 15 years ago. The volume was not small. Their DB model was atrocious. I looked at their wide tables, with tons of columns, not one FK in sight. It was horrible and broke every single rule I've learned in DB design theory, and in practice. It was so bad that I was considering leaving. Luckily I did not leave. With time, I learned how the system worked. I made assumptions about how they used the DB which were not correct. It was not really used as a relational database. It was used as a journal of transactions that served as a backup to a system that actually held everything in RAM. Everything worked beautifully together. The system was fast and robust. I learned a lot at that job. Most importantly, I learned not to pass judgement too quickly.