I saw it put really well the other day. Any software has in general a set number of bugs per lines of code. Something like Debian the number of bugs goes down after release as only bugfixes occur, while anything constantly moving like a rolling release, is certain to grow in number of bugs as the less tested newer software (which generally includes more loc) is pushed. There are tradeoffs to both methods, and edge cases of course.
These days id prefer a developer produce negative lines of code without breaking anything.
As experience tells me, every program contains at least one bug.
Experience also tells me, that you can remove the buggy line of code and the program will still not work as intended.
From this follows, that every program can be reduced to a single line of code that doesn’t work as intended.
Wasn’t there a kernel release a few years back that actually resulted in less code? Or at least at some huge part?
I saw it put really well the other day. Any software has in general a set number of bugs per lines of code. Something like Debian the number of bugs goes down after release as only bugfixes occur, while anything constantly moving like a rolling release, is certain to grow in number of bugs as the less tested newer software (which generally includes more loc) is pushed. There are tradeoffs to both methods, and edge cases of course.