Well said! I can only imagine this problem is magnified in hardware, where the feedback loop from detecting a failure to fixing it must be very expensive, hence preventing it by designing "cleanly" is even more valuable.
I love the idea of "building quality" into the system and not treating it as a second order problem to solve. There is definitely a role for monitoring, but if you don't have the system built right, you can't - ever - fix it systemically.
100% agree. Setting up the structure _inside_ of the stack is a lot of work, but the impact on quality is excellent. And if it's supported by solid monitoring/alerting/testing, even better.
There is a direct analogy with HW development. Any high complex tool has worse reliability than a simple "clean" design by the theory.
Quality and passion for the highest possible simplicity must be cn integral part of each true developer whatever product is made.
Well said! I can only imagine this problem is magnified in hardware, where the feedback loop from detecting a failure to fixing it must be very expensive, hence preventing it by designing "cleanly" is even more valuable.
I love the idea of "building quality" into the system and not treating it as a second order problem to solve. There is definitely a role for monitoring, but if you don't have the system built right, you can't - ever - fix it systemically.
100% agree. Setting up the structure _inside_ of the stack is a lot of work, but the impact on quality is excellent. And if it's supported by solid monitoring/alerting/testing, even better.