Hacker News Reader

What would happen if we applied this for 4 years on an existing codebase, such as the Linux Kernel?

People would fix the low hanging fruit and leave the hard/hairy parts that need work the most.

True - but then eventually, everyone would have to focus on the hard/hairy parts, no?

Yup. That doesn't mean that it's not worth taking care of the easy bits first.

Most likely support for obscure devices would disappear.

To be fair, that's already starting to happen[0]. do you think it would happen faster than it is today?