30 yrs in the business and I never run into NULL semantics issues... because I virtually never use NULL (!) my experience: it's far better to explicitly reserv

30 yrs in the business and I never run into NULL semantics issues... because I virtually never use NULL (!) my experience: it's far better to explicitly reserve a value for NULL. In fact, I go the other way and ask when NULL is worth adding? example: in your datatype NULL-values exist but there's no easy way to represent them otherwise, i.e. you can't reserve a value because they're truly all used by real data. example: you have a slew of BOOLEAN columns and NULL is a legitimate third value in the dataset. The alternative (char(1)) can be a lot of overhead vs the bitmask optimizations that most SQL engines offer (but test this!!!) hope this helps! p.s. having studied the history and known some of the people responsible... if we gave them a time machine, they'd first kill Hitler... and then undo this decision.