He notes several aspects of his development process and architectural decisions that contributed to qmail's success. Interestingly, many of his comments relate to his use of concepts in his C code that are usually associated with higher-level languages. Here he complains about C's dangerous handling of integers:
Another surprise for the programmer is that y can be much smaller than x after y = x + 1. This happens if x is the largest representable integer, typically 2^31 − 1; y will then be the smallest representable integer, typically −2^31 .
...
Most programming environments are meta-engineered to make typical software easier to write. They should instead be meta-engineered to make incorrect software harder to write. An operation that is not exactly what I normally want should take more work to express than an operation that is exactly what I normally want.
Ruby does get many details like this right, and good object-oriented practices can help too. For example, I kept thinking of the Tiny Types pattern while reading this.
I'm reading this a bit at a time and still parsing what Bernstein's ideas mean to me, but the segments on spending some performance dollars to buy quality abstractions were refreshing to see from a developer who's met with measurable success.