Goodhart's Law: When a measure becomes a target, it ceases to be a good measure.
I worked at a company where management measured coder's productivity by LoC (lines of code). If I wrote a function in 30 lines, and Jeff wrote it in 300 lines, Jeff was considered ten times more productive.
The week before the LoC measurement, every coder on the project #include'd the root database header file. That header file included about 10,000 lines of database scheme definitions, macros, inline functions, forward references, etc. And they included it in every single code file. So the BlipVert() function library, which had 60 source files and was about 4,000 lines of code becameover half a million lines of code.
Compile times quintupled, as did memory usage, and performance ground to a halt.
Management panicked, and had a tiger team assigned to find the issue. A week later, they reported they had "performance tuned" the system, and were awarded bonuses for... removing the #includes they'd added a month earlier.
I worked for one company where we would have at least six months of meetings before any change could be made for testing. There was so much documentation involved that coding was the least of my actual work. I did manage to enjoy that job anyway.
I quit and saved the company infinity dollars! 💸
Goodhart's Law: When a measure becomes a target, it ceases to be a good measure.
I worked at a company where management measured coder's productivity by LoC (lines of code). If I wrote a function in 30 lines, and Jeff wrote it in 300 lines, Jeff was considered ten times more productive.
The week before the LoC measurement, every coder on the project #include'd the root database header file. That header file included about 10,000 lines of database scheme definitions, macros, inline functions, forward references, etc. And they included it in every single code file. So the BlipVert() function library, which had 60 source files and was about 4,000 lines of code becameover half a million lines of code.
Compile times quintupled, as did memory usage, and performance ground to a halt.
Management panicked, and had a tiger team assigned to find the issue. A week later, they reported they had "performance tuned" the system, and were awarded bonuses for... removing the #includes they'd added a month earlier.
I worked for one company where we would have at least six months of meetings before any change could be made for testing. There was so much documentation involved that coding was the least of my actual work. I did manage to enjoy that job anyway.
Like Bob always say, this meeting could be an email.