I saw this comment:
[applications] with the most data-related problems were those using natural keys.
I want more supporting evidence of this, as the comment left much to imagination.
It suggests that practice of using natural keys creates data-related problems, but does not specify what goes wrong... does data get corrupt? out of sync? becomes erroneous, lost, damaged? hard to query?
What are the data problems that happen when database is designed with natural keys opposed to using surrogate keys? How can those type of problems can be prevented when using surrogate keys?