Changing the paradigm on linting
I've spend some time today thinking about linting. With AI developers, just like humans, I want to provide as much automation as I can. GitHub has AI code reviews now and I want to explore that as much as possible. I got thinking about linting. The purpose of linting for human developers is to reduce the cognitive load on the developer who comes after you; they understand the naming conventions, the code is easy to read, there is consistency in how we write code. But do we need that in the same way for AI developers? Sure, maybe its helpful, but I've already got StyleCop running and feeding back as code is written. I'm thinking about how we reduce the cognitive load on AI developers and improving their DevX. I think that the shift is to move this to before coding starts, providing better context, better knowledge, better coding rules. Human developers want to take the ticket and have the flexibility of solving the problem creatively, but AI developers want more guides, more examples and more help. I will ponder on this more.