Sweden’s and Diyanet’s methods, worldwide
The correction lists
The method reproduces Diyanet’s computation; what it cannot reproduce is the
arithmetic of Diyanet’s program, which puts a few times
(0.018 % in Sweden) on the other side of a minute’s rounding point — a fraction of a second
away, independently from place to place and day to day.
data/corrections/sweden.json lists those times (point, date, prayer, a
minute either way: 270 over
1 Jan 2024 → 31 Dec 2027) and timesFor applies them. The list covers the
dates the timetables do, and certification holds over those dates only:
outside them timesFor answers with status: 'method-only' (or refuses,
with strict). Diyanet’s worldwide list (data/corrections/diyanet.json)
is the same kind: 9,101 entries against
71.6 million times. Each year, once the next year’s timetables are
out:
npm run corrections && for a in diyanet morocco algeria indonesia; do npm run corrections -- $a; done && npm run certify
What Diyanet’s worldwide tables add to the Swedish method
- Isha by country: 16° (the Congress’s) in most of Europe, Russia, Ukraine, Belarus, Kazakhstan, Mongolia, China, the United States and Canada; 17° elsewhere. The high-latitude rules are worked at 16° either way.
- The takdir at midsummer: where the dawn forms all summer but midsummer’s true Isha falls later than the share carried from that night, midsummer is the takdir’s day, with the twenty-minute ramps (from about 44.4° north).
- Türkiye’s districts: each prayer is computed at a point of its own in
the district, and the points move by a few thousandths of a degree from one
year’s tables to the next (
data/places/diyanetTurkeyPoints.json, a point per prayer per year; Northern Cyprus the same). - Past local midnight: where the first true dawn after the summer gap falls before local midnight on the table’s clock (far east of a zone’s meridian, as at Aihui in China), there is no transition: the gap’s last day is printed twenty minutes after the takdir and the true times follow.
- Tables on their own clocks: some of Diyanet’s tables are printed on a
clock other than their place’s zone — Alberta’s keep summer time from 2026
on; Mexico’s had it again in 2025; Queensland’s and Western Australia’s
have it in 2024, 2025 and 2027; Paraguay’s keep the summer time the country
gave up in October 2024 until the end of 2025; Xinjiang’s are Beijing time;
western Brazil’s, Armenia’s and some of Russia’s and Mongolia’s are a zone
or two away from their own; Alchevsk’s are Moscow time with summer time; Cuba’s go back to
winter time in November 2026, which Cuba itself put off to November 2027.
For those the table’s clock is read off the table itself (
clockindata/places/diyanetWorldPoints.json); the instants are the method’s. - The day’s clock: a table prints the whole day in the clock kept at noon, so on the morning the clocks go forward a Fajr before the change is printed in the new time.