The continental US alone spans four time zones — Eastern, Central, Mountain, and Pacific — and for decades Indiana was the textbook example of how confusing that can get: until 2006, most of the state didn't observe daylight saving at all while its neighbors did, so crossing a county line could mean the clock jumped by an hour or stayed put depending on which side of an old boundary you were on.
Why Indiana became a cautionary tale
Before a 2006 state law standardized the rules, dozens of Indiana counties independently chose whether to follow Eastern or Central time and whether to observe daylight saving, based on local custom rather than any federal mandate. Airlines and TV schedules had to account for a state that, on paper, ran on multiple inconsistent clocks at once.
Offset relative to UTC
Every time zone is defined by an offset from UTC — US Eastern Time is UTC-5 in winter and UTC-4 during daylight saving. Offsets aren't always whole hours either: India runs on UTC+5:30, and Nepal on UTC+5:45.
Why time zone conversion is hard in practice
Daylight-saving rules are set by individual states, provinces, or countries and can change with little notice — Arizona famously opts out of DST entirely (except the Navajo Nation within it, which observes it). Reliable time conversion needs a constantly updated time zone database, like the IANA Time Zone Database, rather than a fixed hourly offset.
Why you'd need this
- Scheduling a call between teams in New York, Denver, and London without anyone showing up an hour off.
- Converting timestamps from server logs into a user's local time.
- Avoiding scheduling errors around daylight-saving transitions.
Not all offsets are whole hours
Although most time zones are offset from UTC by a whole number of hours, there are exceptions: India uses UTC+5:30, Nepal uses UTC+5:45, and some regions of Australia use UTC+9:30 or UTC+10:30. Code that assumes an offset is always a multiple of 60 minutes should be avoided — this is exactly why reliable libraries work with IANA timezone names rather than arbitrary numeric offsets.