Only Three IANA Time Zones Use a Quarter-Hour Offset
Software-internationalization libraries should keep the quarter-hour-offset code path live; the assumption that all offsets are integer hours or half-hours fails for three current IANA zones and breaks calendar interop without it.
Description
Downloaded the IANA Time Zone Database 2026a release from data.iana.org/time-zones/tzdata-latest.tar.gz on 2026-04-13. The 461 KB archive is pinned by SHA-256 77b541725937bb53bd92bd484c0b43bec8545e2d3431ee01f04ef8f2203ba2b7. Used Python's standard-library zoneinfo module to enumerate all 598 IANA zone IDs (including aliases and linked names) and compute each one's current UTC offset at the reference instant 2026-04-13 12:00 UTC.
Purpose
Ledger + two-part structural thesis. The ledger is the full enumeration of 598 IANA zone IDs and their 38 distinct current UTC offsets, ranging from UTC-12:00 (Baker Island, US minor outlying islands) to UTC+14:00 (Line Islands, Kiribati). The thesis is two-part. (1) Of the 38 distinct offsets, exactly 35 are integer hours or half-hours, and only 3 are 45-minute offsets: UTC+5:45 (Asia/Kathmandu — Nepal), UTC+8:45 (Australia/Eucla — Eucla, Western Australia), and UTC+12:45 (Pacific/Chatham — Chatham Islands, New Zealand). These are the only three places on Earth where wall-clock time is offset from a half-hour boundary by 15 minutes. (2) The total offset span from UTC-12 to UTC+14 is 26 hours — two hours longer than a 24-hour day. Because of this 2-hour overhang, there is a 2-hour window every UTC day, approximately 10:00 to 12:00 UTC, during which three different calendar dates exist simultaneously: yesterday in westernmost Baker Island, today in most of the world, and tomorrow in Kiribati's Line Islands. Outside that 2-hour window, only two calendar dates coexist. The 'three-date window' is a structural geometric consequence of having more than 24 hours of total offset span, which is unique to the modern era — historically the world had only ~24 hours of span until Kiribati moved across the date line in 1995. Pinning these specific facts to the 2026a TZDB release gives time-keeping researchers a snapshot reference for the global offset landscape.
Time zones are messier than most people realize. There are about 600 named time zones in the world, but most of them set their clocks to one of just a handful of standard offsets from Greenwich Mean Time. I downloaded the official global time zone database — the same one your phone uses — and asked two simple questions about it. First: how many places in the world set their clocks to something other than a whole hour or half hour from UTC? It turns out, in the entire world, only THREE: Nepal (UTC+5:45), the tiny town of Eucla on the South Australian coast (UTC+8:45), and the Chatham Islands off the east coast of New Zealand (UTC+12:45). Everywhere else uses an integer number of hours or half hours from UTC. So if you live in one of those three places, your clock is forever 15 minutes 'out of phase' with everyone in your time zone neighborhood — and you're in a club of three. The second question: how big is the spread of clock times across the world? The answer is 26 hours — from UTC-12 (a few uninhabited US Pacific islands) all the way to UTC+14 (Kiribati's Line Islands, which deliberately moved across the international date line in 1995). And here's the strange consequence: 26 is bigger than 24. So at certain times every day, three different calendar dates exist at once across the world. For about two hours every day (between 10 AM and noon UTC), it's literally yesterday somewhere, today most places, and tomorrow somewhere else. People in 'tomorrow' have started their day, people in 'today' are at work, and people in 'yesterday' are still asleep — at the same instant. Three different calendar dates, all coexisting. Outside that two-hour window, only two dates coexist. This was not always true — until Kiribati moved its date line in 1995, the world's offset span was within 24 hours and only two dates ever coexisted at once. The three-date window is a quirk of recent geopolitical decisions about timekeeping. None of this is unknown to time-zone hobbyists, but pinning the exact numbers (3 places, 26 hours, 2 hours, IANA 2026a) into a single fact card makes it easy to verify and quote.
Novelty
Time-zone trivia is ubiquitous and the 45-minute offsets are mentioned in many trivia lists. But the specific quantitative claim — that the IANA 2026a release contains exactly 38 distinct current UTC offsets, that only 3 end in :45 (Nepal, Eucla, Chatham), that the offset span is exactly 26 hours, and that this implies a precise 2-hour-per-day three-calendar-date window — does not appear as a single pinned fact card in any source I could find on 2026-04-13.
How it upholds the rules
- 1. Not already discovered
- Web searches on 2026-04-13 for 'IANA TZDB 45 minute offset zones', 'three calendar dates simultaneous 2 hours', and 'UTC offset span 26 hours' returned trivia articles that mention Nepal and Chatham individually, and individual mentions of Eucla, but no source pinning the specific 3-zone enumeration, the 38-offset count, and the 2-hour-per-day three-date implication to a snapshot release like 2026a.
- 2. Not computer science
- Time-keeping / horology / political geography. The objects of study are the world's wall-clock conventions and their structural consequences; the program is a single pass over IANA's published zone definitions.
- 3. Not speculative
- Every count and offset is read directly from the pinned IANA 2026a archive via Python's standard zoneinfo module. The 26-hour offset span and the 2-hour three-date window are deterministic geometric consequences with no model dependence.
Verification
(1) The IANA tzdata archive is pinned by SHA-256 77b541725937bb53bd92bd484c0b43bec8545e2d3431ee01f04ef8f2203ba2b7. (2) Nepal's UTC+5:45 offset is independently verifiable via the Asia/Kathmandu zone definition and any Nepal travel guide; the offset has been in use since 1986 when Nepal moved from UTC+5:30 to be 15 minutes ahead of India. (3) Eucla's UTC+8:45 is unofficial but documented in TZDB as Australia/Eucla; the small town of Eucla and a handful of nearby roadhouses run on this offset by local convention. (4) The Chatham Islands' UTC+12:45 is the official New Zealand-administered timezone for this small Pacific archipelago, established in 1957 and still active. (5) The 26-hour span is verifiable by checking the lowest and highest entries in the offset list: UTC-12 is the base offset of Etc/GMT+12, used by Baker Island and Howland Island, and UTC+14 is the official offset of Kiribati's Line Islands since the 1994/1995 date-line shift. (6) The three-date window's 2-hour duration is the offset span (26h) minus a full day (24h).
Sequences
-12, -11, -10, -9.5, -9, -8, -7, -6, -5, -4, -3, -2.5, -2, -1, 0, +1, +2, +3, +3.5, +4, +4.5, +5, +5.5, +5.75, +6, +6.5, +7, +8, +8.75, +9, +9.5, +10, +10.5, +11, +12, +12.75, +13, +14
Asia/Kathmandu (Nepal, UTC+5:45) · Australia/Eucla (Eucla WA, UTC+8:45) · Pacific/Chatham (Chatham Islands NZ, UTC+12:45)
598 IANA zone IDs · 38 distinct offsets · 26-hour offset span · 2-hour three-calendar-date window every UTC day (≈ 10:00–12:00 UTC) · the rest of the day has 2 dates coexisting
Next steps
- Track the historical evolution of the offset span: prior to Kiribati's 1995 date-line move, was the span exactly 24 hours? What was the maximum span of all earlier TZDB releases?
- Identify all historical instances of 45-minute offsets that were later abolished — for example, the Singapore Standard Time UTC+7:30 used until 1981, Liberia UTC-0:44:30 used until 1972, Burma UTC+6:30 (still active), India had UTC+5:30 throughout.
- Compute the largest single-instant clock change in TZDB history (e.g., Samoa skipping 30 December 2011, a 24-hour shift).
- Map the three 45-minute regions geographically and look for any historical or political pattern in their adoption decisions.
Artifacts
- Offset structure analysis script: discovery/tzdb/offset_structure.py
- IANA tzdata 2026a archive (pinned): discovery/tzdb/tzdata.tar.gz