The most common date format used internally is “seconds since January 1st, 1970”.
In early 2038, the number of seconds will reach 2^31 which is the biggest number that fits in a certain (also very common) data type. Numbers bigger than that will be interpreted as negative, so instead of January 2038 it will be in December 1901 or so.
What is the basis for the 2038 problem?
The most common date format used internally is “seconds since January 1st, 1970”.
In early 2038, the number of seconds will reach 2^31 which is the biggest number that fits in a certain (also very common) data type. Numbers bigger than that will be interpreted as negative, so instead of January 2038 it will be in December 1901 or so.
Huh interesting. Why 2^31? I thought it was done in things like 2^32. We could have pushed this to 2106.
Signed integers. The number indeed goes to 2^32 but the second half is reserved for negative numbers.
With 8 bit numbers for simplicity:
0 means 0.
127 means 127 (last number before 2^(7)).
128 means -128.
255 means -1.