I think it is a typical "depends" situation. I also favour writing new things, but in practise there are many situations where this would be the act that makes

I think it is a typical "depends" situation. I also favour writing new things, but in practise there are many situations where this would be the act that makes you look like you enjoy pain. E.g. if you have legacy systems that cover and handle edge cases you didn't even know existed. If you build your new thing you are going to learn about them. Probably in production. Suddenly that fresh feeling of writing a new thing will be replaced by another, considerably worse feeling. It really depends on the problem your old thing is solving. If it is a simple problem, then a new thing can be a no-brainer. If that old tool is however interconnected with other things in a hundred places, good luck. If the problem you are solving is complex but very contained (e.g. just a tool that takes document X and transforms it into document Y), easy — wrire a new one. But I know many operational people who are genius at understanding their systems and tools, but are not good at developing software.