>have abandoned decades of experience showing how to create reliable, robust, reusable code that is both great the customer, fast to iterate on, and only a tiny tiny bit slower to write.
This era you think existed… didn’t. All of the trash code from 10-20-30 years ago is long gone. The only stuff that remains is the <1% that was simultaneously written with quality in mind while also remaining useful.
The quality of software is absolutely irrelevant if you build the wrong thing, so you need to iterate quickly unless you’re lucky enough to know exactly what needs to be built. And even in that case, you need to weigh the engineering cost against a cost of a failure and the cost of expensive maintenance.
Many times, 100 lines of very use-case specific python that can’t be re-used is absolutely the correct engineering call. It might not be what you like as a programmer, but you’re not paid to navel-gaze.
This era you think existed… didn’t. All of the trash code from 10-20-30 years ago is long gone. The only stuff that remains is the <1% that was simultaneously written with quality in mind while also remaining useful.
The quality of software is absolutely irrelevant if you build the wrong thing, so you need to iterate quickly unless you’re lucky enough to know exactly what needs to be built. And even in that case, you need to weigh the engineering cost against a cost of a failure and the cost of expensive maintenance.
Many times, 100 lines of very use-case specific python that can’t be re-used is absolutely the correct engineering call. It might not be what you like as a programmer, but you’re not paid to navel-gaze.