>It's much easier to learn good habits than to unlearn bad ones.
How do you know they are 'good habits'. I have seen countless years of bad practices lauded internally as amazing/the etalon weight when it comes to code quality. In reality most of them were textbook examples of what should not be done. When you get folks without any previous experience, there's no one to question the status or the authority. If they learn/wisen up, they are likely to leave.
Recent grads tend to be more evidence-oriented than people with experience, IME. They'll e.g. benchmark something to see whether it's faster rather than going by reputation alone.
Hmm - that's quite nice/reassuring, although not my experience. Benchmarking, OTOH, is notoriously hard, esp. the microbenchmark type. The old: lies, damn lies, statistics, (micro)benchmarks.
How do you know they are 'good habits'. I have seen countless years of bad practices lauded internally as amazing/the etalon weight when it comes to code quality. In reality most of them were textbook examples of what should not be done. When you get folks without any previous experience, there's no one to question the status or the authority. If they learn/wisen up, they are likely to leave.