Talk:Antimeridian

From Geohashing
Revision as of 19:01, 1 August 2020 by January First-of-May (talk | contribs)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)

IIRC, technically, the graticules on different sides of the Antimeridian follow different versions of the 30W rule, so in a theoretical zero offset on either side, only 180 hashes would fall here (unless both offsets are zero, which is doubly unlikely).

Yes, this also means that any theoretical (e.g.) Samoan geohashers would get their hashes sometime early morning of next day. But the alternative was allowing graticules like Matavai, Samoa to have two valid hashes on the same day. --January First-of-May

You are right, I did not consider this. Ironically, I raised this issue in a different context a few years ago. Currently, all implementations that I know about use the Antimeridian as the date line, which of course is not true.
It would be fair to Geohashers east of the Antimeridian but west of the date line if they would also be protected by the W30 rule. They should be considered east of W30, not west of W30. Of course that would lead to two geohashes in the same graticule on the same day in graticules where the date line cuts through, so a date-lat-lon combination would not be unambiguous, which of course is unacceptable.
One way to tackle this would be to assign a value to all graticules if they should be considered east or west of the date line. That way there is only one valid geohash each day. Your example, Matavai, Samoa, should be considered west of the date line, since all of it's land is west of the date line. On the other hand, Wales, Alaska, should be considered east of the date line since most of it's land is east of the date line, even though a tiny sliver of land is west of the date line.
Of course, this would be a nightmare to go through all graticules, implement it into a program, and nobody I know has done this so far. I know it is lazy, but it is probably not worth worrying about this problem since there aren't any geohashers in these areas anyway, at least currently.
I'll change the text on the page to say that it's 180 locations, since that is the interpretation that the implementations currently follow. --Fippe (talk) 16:59, 1 August 2020 (UTC)
I agree! Arguably this is what Randall meant in the first place when he did the "after noon local time" version.
I don't think there are any graticules other than Wales, Alaska with land areas on both sides of the line. I agree that implementation would be a mess, though (and liable to further changes if some other Pacific country desides to change which side it's on).
It also might be a little unfair to geohashers in looks at the map Shemya, Alaska, who would no longer get their weekday hashes nearly a day in advance, though we could probably ignore them anyway because there's only like a dozen people living in this entire area. --January First-of-May (talk) 18:59, 1 August 2020 (UTC)