Hacker News Reader

I find the stack exchange interesting, but I still think its post-hoc reasoning. I would prefer one of the people who worked on these systems to say it, than have a plausible, but none-the-less third hand take published as "the answer"

View on HN

I didn't wok on one of the US government systems, but much of my career was in COBOL compilers (writing run time systems, and checking against the ANSI standards). COBOL would not default to 1875 for a corrupt date.

> COBOL would not default to 1875 for a corrupt date

But COBOL doesn't have a date type.

So isn't its behaviour defined by how the system has been implemented ?

We already have "the answer", it's obviously bad or exceptional data. The question "is it fraud" can't be answered by looking narrowly at that data.

But the idea that no one has noticed the problem and investigated it (including possible fraud) is very unlikely.

Agreed. And, the persistence of bad data like this isn't uncommon. We had a bunch of fields in the SQL registry system I work in set to the literal 4 character ascii string "null" because of a coding error over field == null by somebody. It did no harm to the attributes of interest at the time. It only had consequences later on when the codebase changed.

Reply on news.ycombinator.com