Difference between revisions of "Talk:Globalhash"

From Geohashing
imported>TrueNarnian
(Better Distribution of Points?)
(Eliminating redlinks)
 
(2 intermediate revisions by one other user not shown)
Line 24: Line 24:
 
=== Australian Globalhash ===
 
=== Australian Globalhash ===
  
The hash for Friday [[2009-03-13 global|2009-03-13]] looks very reachable for a globalhash. Any chance of a good Aussie expedition? --[[User:Woodveil|Woodveil]] 14:55, 12 March 2009 (UTC)
+
The hash for Friday {{egl|2009-03-13|global|text=2009-03-13}} looks very reachable for a globalhash. Any chance of a good Aussie expedition? --[[User:Woodveil|Woodveil]] 14:55, 12 March 2009 (UTC)
  
 
:Hmmm - would love to try this one if I had nothing on today. It's about five hours north of Sydney, and probably achievable. I don;t know of any active geohashers in the northern part of our state though... --[[User:CJ|CJ]] 15:04, 12 March 2009 (UTC)
 
:Hmmm - would love to try this one if I had nothing on today. It's about five hours north of Sydney, and probably achievable. I don;t know of any active geohashers in the northern part of our state though... --[[User:CJ|CJ]] 15:04, 12 March 2009 (UTC)
Line 240: Line 240:
 
::::Agreed with Benjw and Sourcerer... hahahaha.  Also, I think you may have misunderstood me [[User:Laurint|Laurint]] - of course I understand there will be more globalhashes within an x-km radius of you if you live near the poles than if you live near the equator.  What I was trying to say back in 2014 was that <b>you</b>, Laurint, live quite near the Pole yourself - so the current arrangement actually benefits you!  If we were to switch to arcsin, your globalhash frequency would decrease (even as mine, down here at latitude 38N, would increase.)  So I was trying to defend you from the implication that you were trying to make your globalhashing job easier. :D -- [[User:OtherJack|OtherJack]] ([[User talk:OtherJack|talk]]) 21:18, 20 January 2017 (UTC)
 
::::Agreed with Benjw and Sourcerer... hahahaha.  Also, I think you may have misunderstood me [[User:Laurint|Laurint]] - of course I understand there will be more globalhashes within an x-km radius of you if you live near the poles than if you live near the equator.  What I was trying to say back in 2014 was that <b>you</b>, Laurint, live quite near the Pole yourself - so the current arrangement actually benefits you!  If we were to switch to arcsin, your globalhash frequency would decrease (even as mine, down here at latitude 38N, would increase.)  So I was trying to defend you from the implication that you were trying to make your globalhashing job easier. :D -- [[User:OtherJack|OtherJack]] ([[User talk:OtherJack|talk]]) 21:18, 20 January 2017 (UTC)
 
::::(Btw it's actually a fun math problem to figure out the latitude separating the places that would "win" from those that would "lose" globalhashes due to the change... turns out it's not 45, but somewhere in the low 50s, around London or maybe Sourcerer's location.  In the current system the chance of a globalhash landing in your local y km N-S by x km E-W box is basically y/(pole-to-pole distance) * x/(circumference of your latitude circle) = y/(piR)*x/(2piRcos(lat)) = xy/(2pi^2 R^2 cos(lat)), where R is earth's radius.  Whereas in the future proposed system, it is the same everywhere, xy/(area of Earth) = xy/(4pi R^2).  So if the rules were to change, one's chances would improve by a factor of (2pi^2 R^2 cos(lat))/(4pi R^2) = (pi/2) cos(lat).  E.g. at latitude 0, your frequency would be multiplied by (pi/2) * 1 = 1.57, so 57% more "likely" to find a global point than you were before the rule change.  But at Laurint's latitude of 60, it would be multiplied by (pi/2) * (1/2) = 0.79, so 21% less likely than before the change.  The crossover is where (pi/2) cos(lat) = 1.00 exactly, so we need cos(lat) = 2/pi.  The google tells me arccos(2/pi) is about 50.4597763 degrees, which lines up with the English Channel or central Germany.) -- [[User:OtherJack|OtherJack]] ([[User talk:OtherJack|talk]]) 21:18, 20 January 2017 (UTC)
 
::::(Btw it's actually a fun math problem to figure out the latitude separating the places that would "win" from those that would "lose" globalhashes due to the change... turns out it's not 45, but somewhere in the low 50s, around London or maybe Sourcerer's location.  In the current system the chance of a globalhash landing in your local y km N-S by x km E-W box is basically y/(pole-to-pole distance) * x/(circumference of your latitude circle) = y/(piR)*x/(2piRcos(lat)) = xy/(2pi^2 R^2 cos(lat)), where R is earth's radius.  Whereas in the future proposed system, it is the same everywhere, xy/(area of Earth) = xy/(4pi R^2).  So if the rules were to change, one's chances would improve by a factor of (2pi^2 R^2 cos(lat))/(4pi R^2) = (pi/2) cos(lat).  E.g. at latitude 0, your frequency would be multiplied by (pi/2) * 1 = 1.57, so 57% more "likely" to find a global point than you were before the rule change.  But at Laurint's latitude of 60, it would be multiplied by (pi/2) * (1/2) = 0.79, so 21% less likely than before the change.  The crossover is where (pi/2) cos(lat) = 1.00 exactly, so we need cos(lat) = 2/pi.  The google tells me arccos(2/pi) is about 50.4597763 degrees, which lines up with the English Channel or central Germany.) -- [[User:OtherJack|OtherJack]] ([[User talk:OtherJack|talk]]) 21:18, 20 January 2017 (UTC)
 
I may be late to the party on this, but I think I have come up with an algorithm that is similar to the existing method, is relatively easy to implement and distributes the globalhash points perfectly evenly (assuming a spherical earth).  For the longitude, use the same method as before (X = 360x - 180), but for the latitude, use Y = sin<sup>-1</sup>(2y - 1).  In other words, take the decimal form of the first half of the hash (as you would normally), double it, subtract 1, and then take the Arc Sine of this number in degrees as the latitude.  What you have essentially done is mapped the hash to the earth's axis.  An initial value of 0 is the South Pole, and 1 is the North Pole.  Assuming any value between 0 and 1 is equally likely, it is thus equally likely to get a point anywhere on this line through the center of the earth.  Then, draw an imaginary line from this point, perpendicular to the earth's axis, to the surface (in the direction of the latitude).  Because of the geometry of a sphere, this means that it is equally likely that the hash will fall on any point on the surface of the globe. -- [[User:TrueNarnian|TrueNarnian]] 16:13, 26 May 2017 (EDT)
 
  
 
== A name for 'Globalocal' points? ==
 
== A name for 'Globalocal' points? ==

Latest revision as of 10:17, 22 March 2020

I love the idea. Just as Ph. or M.S. Fogg would do. --Tom 17:32, 3 September 2008 (UTC)

I just need someone to help code this; I can edit code a bit, but I wouldn't know where to start with this. I have no real web coding expertise. -Wmcduff 17:41, 3 September 2008 (UTC)

I keep an eye on the Globalhash, just in case... today (16th of February) it falls in Libya. If you zoom in on satellite view about half way and look a little north there are a lot of black circles. Does anyone know what they are? http://www.drwilco.net/globalhash/map.html. --Kate 13:23, 16 February 2009 (UTC)

They look like center pivot irrigation systems, but could be anything. Some of them appear green at certain times. On the ones to the left you can make out the device that circles these circles. -- relet 13:32, 16 February 2009 (UTC)
S-i-g-h-t-e-d-P-e-o-p-l-e-S-u-c-k --Ilpadre 15:16, 16 February 2009 (UTC)

2014-07-10 in Vancouver graticule by air?

See the discussion at Talk:Vancouver, British Columbia...

Interesting globalhash locations

Monday's (2012-10-22) is smack in the middle of lake Tonle Sap, a huge, unusual lake in Cambodia that fills up every rainy season with water backing up its outlet from the nearby river, and then drains back out the same outlet during the dry season. As I just learned from the above link. -- OtherJack 15:44, 19 October 2012 (EDT)

And this Sunday's (2013-02-03) in Southeast Alaska must be the most visually stunning globalhash point ever. (If you zoom in further you can see it's RIGHT on the coastline there...) -- OtherJack 16:37, 1 February 2013 (EST)
Tuesday's (2013-10-01_global) is in TRANSYLVANIA. And to boot, it's 30m from a road, in a scenic valley in the Carpathians. You can see the globalhashpoint in google street view! -- OtherJack (talk) 22:59, 30 September 2013 (EDT)

2012-09-05

It's in the Clearwater National Forest, north Idaho. I am seriously thinking of renting a car and driving out there, leaving later this afternoon, if it looks hikable. Anyone else thinking of going and doing this? I made a stump expedition page at 2012-09-05 global. -- OtherJack 16:09, 4 September 2012 (EDT)

No hiking trails, no time for very long drive + very long bushwack, so no go. See the page for sad details... -- OtherJack 18:22, 4 September 2012 (EDT)

Australian Globalhash

The hash for Friday 2009-03-13 looks very reachable for a globalhash. Any chance of a good Aussie expedition? --Woodveil 14:55, 12 March 2009 (UTC)

Hmmm - would love to try this one if I had nothing on today. It's about five hours north of Sydney, and probably achievable. I don;t know of any active geohashers in the northern part of our state though... --CJ 15:04, 12 March 2009 (UTC)
If I read right, the globalhash in question is for Friday. So if you have 5 hours of spare time tomorrow...--Woodveil 15:09, 12 March 2009 (UTC)
But in Sydney, it's tomorrow already, I think! (And by the way, is Friday 13th really the best day to be globalhashing?) :-) -- Benjw 15:21, 12 March 2009 (UTC)
Yes - it certainly is Friday here already! We live in the future here in Australia (well, compared to you at least) --CJ 16:13, 12 March 2009 (UTC)
That's a great location, but ... are you using 2009-03-12 DOW and 2009-03-13 as the date to generate the coordinates? How does this work for east-of 30W?
I check the globalhash daily using the Online Globalhashing Tool, as featured on the Globalhash page. I personally don't know what numbers it uses, but it always shows me the globalhash a day in advance. From what I can tell, it gives the entire earth the same hash for a given day, using the Dow opening from the day before. So for anyone East of 30W, the globalhash for a given day uses the same Dow opening as that day's geohash (i.e. the previous day's Dow opening). But West of 30W, the globalhash also uses the previous day's Dow opening even though the geohash uses the Dow opening from that morning. I hope that makes sense to anyone else... Or could anyone explain it better? --Woodveil 16:07, 12 March 2009 (UTC)
Okay, I see it now, it always uses the date before what you enter, (which works out to the same day for the eastern hemisphere,) and the Dow opening for that day. If you check for a Monday, it will presumably use the preceding Friday's open, and Sunday's date. That way it can get the same globalhash for the whole world, and with a meaningful amount of time to reach it. -- Jevanyn 17:52, 12 March 2009 (UTC)

Globalhashes for Easter weekend

The Online Globalhashing Tool appears to be unaware that the Dow is closed for Good Friday (2009-04-10). So, although I can see the 4/10 globalhash (eastern Russia), I can't see the points for Saturday, Sunday, or Monday. --starbird 01:55, 10 April 2009 (UTC)

I use this tool (fill in your exact lat and long in the address) which has them ok. They are:
  • Today: N 65°14.990 E 165°29.954. Graticule: Unnamed, probably Russia
  • Saturday: N 2°34.982 E 51°04.899. Graticule: Unnamed, probably Northern Africa
  • Sunday: S 24°21.339 E 150°56.262. Graticule: Biloela, Australia
  • Monday: S 76°43.414 E 122°49.570. Graticule: Unnamed, probably Antarctica
What I am mildly surprised about is that another day's hashes haven't been released. Is Easter Monday not a holiday in the USA? -- Benjw 07:35, 10 April 2009 (UTC)
Thanks, Benjw. No, some US businesses (and the financial markets) are closed on Good Friday, but the following Monday is not generally observed.
Sunday's point looks like a second chance for Brisbane and some of the hardcore crazies further south (see the 2009-03-01 globalhash). Saturday's point off Somalia is more intruiging, though. What kind of ribbon do you get for meeting REAL PIRATES at a water hash? --starbird 08:57, 10 April 2009 (UTC)
Trail of blood consolation prize, I should think! -- Benjw 11:22, 10 April 2009 (UTC)
At least. There are several naval ships in the area as well (US and India off the top of my head) and the latest hijacking in the area was pretty close to those coordinates. -- Jevanyn 15:09, 10 April 2009 (UTC)

Upcoming Locations

* May 11, 2009 offers an incredible opportunity for the intrepid Globalhasher. Located, as near as I can tell, exactly on the Clark Fork of Red Creek in Yosemite National Park, this location sits approximately seven miles (as the crow flies) from Curry Village, and about a mile and a half from the nearest hiking trail. If you choose to accept this mission, you will collect (at least) three achievements: Globalhash, Water Geohash, and MNIMB Geohash. Good luck!

Somebody's calculations are a little out. The globalhash for Monday 11 May 2009 is at N 87°18.227', W 124°10.674'. -- Benjw 14:31, 8 May 2009 (UTC)

Hmm. Interesting. I was going from the drwilco map generator. Odd. Bill^2

It hasn't picked up the 8 May Dow opening yet, so it's using a DJIA of 0.0 for all three hashes this weekend. --starbird 15:27, 8 May 2009 (UTC)
I'm thinking that drwilco is completely unreliable at this point, as it still has today's location where I referenced it, and a quick check of tomorrow's shows that to be wrong as well. Anybody know how to reach the person who maintains it? Bill^2
That would explain it. Does it not warn you when it does that?  :-s I use the tool I mentioned in a previous conversation on this page; it seems to be more reliable. -- Benjw 15:31, 8 May 2009 (UTC)
I don't think it realizes. If you ask it for Tuesday's globalhash, it correctly objects. --starbird 15:45, 8 May 2009 (UTC)
It seems to believe 0 is a totally reasonable DJIA value. Although I don't object to that point of view, it's not very likely that that will ever occur soon. About Small Hash Inquiry Tool: That one will always know the globalhash at the very same time peeron has the standard hashes. It gets the DJIA from peeron but does the actual calculations itself. Unless the DJIA returned from there is wrong, the coordinates should be right. --Ekorren 15:52, 8 May 2009 (UTC)

May 11, 2009

Awesome location today: it's in a creek in Yosemite National Park. It's about 10 miles from Curry Village, which is inside the park. For a non-GPSer like me, this looks so reachable. Shows what I know about hiking, etc. :-)

Interesting how the Google Map for terrain has the trails marked, not the roadmap. -- Jevanyn 17:10, 11 May 2009 (UTC)

The drwilco calculator is hosed. The Monday globalhash is in 87, -124, not in Yosemite. See discussion above. It's wrong for Tuesday, too -- the actual point is just a couple of miles off the Vietnamese coast near Da Nang, not in Siberia. --starbird 17:27, 11 May 2009 (UTC)
Oh I'm so glad that is still a bug, but the bug has to be fixed, or that globalhash look up program taken down. I saw it on Friday, but when I made a page for it, I was told by multiple people that it was wrong, that it was at the north pole. When I saw this morning that it still showed in the wrong place, I was afraid that the people who said it was wrong had been mistaken, and that I would have to hate them forever for making me miss a globalhash. -Robyn 17:38, 11 May 2009 (UTC)
I agree, that would be very sad. I plan to make liberal use of the "Toggle debugging info" link in the lower left corner of the calculator page for a while. --starbird 17:50, 11 May 2009 (UTC)
How will that tell me if it is lying? -24.83.1.77 18:28, 11 May 2009 (UTC)
A good test is to check the day's coordinates for places east of the 30W line, and see if they vaguely match up with the globalhash coords. For instance, on May 11, the 30W latitude fraction was 0.98502, which ties up well with a globalhash at 87N, and not at all well with one at 38N, which is roughly where Yosemite is. -- Benjw 18:49, 11 May 2009 (UTC)
If you toggle the debugging info on on the map, it should display a DJIA number used for the day. This weekend, it somehow failed to retrieve the actual DJIA, and thus shows none and calculates with zero. Same thing happens always for a few minutes after DJIA announcement. --Ekorren 19:08, 11 May 2009 (UTC)
It still seems like it's broken; still showing 0 for the DJIA Crazycaveman 15:43, 13 May 2009 (UTC)

2009-05-19

Globalhash for tuesday is in the USA, and even on land, but...

It's on obviously very exclusive private property. You can book daytrips to the island, but I wouldn't expect them to allow you to leave the beaten track and detour into what is probably a swamp. If anyone would like to go for closest failure yet, this might be the day, though ;-)

Here's the homepage of the facilily: Little St. Simons Island lodge

"An enticing array of outdoor activities" could certainly include geohashing! --Thomcat 14:01, 18 May 2009 (UTC)
The hashpoint seems to be in the middle of Myrtle Pond. They do birdwatching there. I doubt if they allow boats on it, but maybe you could smuggle in an inflatable kayak. --starbird 14:18, 18 May 2009 (UTC)

Just for fun, I just tried to find out just how exclusive they are. Prices seem to start at 650$ for one night. It might be cheaper to visit a globalhash on another continent, then ;) --Ekorren 19:37, 18 May 2009 (UTC)

Could someone post a link to a working globalhash app? I'm still looking at this one and being confused. -Robyn 19:49, 18 May 2009 (UTC)
The Small Hash Inquiry Tool (which is linked on the Globalhash page) includes the globalhash as well as your local one(s). Thanks, Ekorren! --starbird 20:02, 18 May 2009 (UTC)
Tuesday's Globalhash is: N 31°16.998 W 81°17.615, located in the Brunswick, Georgia graticule. --Ekorren 20:07, 18 May 2009 (UTC)
The resort also offers day trips, which are presumably a little less expensive. --starbird 20:02, 18 May 2009 (UTC)

2009-05-30

This one is probably the nearest globalhash to a really active graticule for a long time... but even considering that that near graticule is Vancouver, it doesn't really look very accessible...

[Globalhash for 2009-05-30]

2009-06-02

Second British Columbia geohash after only a few days! On land this time... but not really more reachable...

[Globalhash for 2009-06-02]

Monday, 2009-06-08: Queensland, Australia

Somewhere between Townsville and Brisbane, on land... Link to map

2009-06-23: Brewarrina graticule, New South Wales, Australia

Another Australia land globalhash, even looks almost accessible on the satellite photo - and still, nobody going?

Map

2009-08-18 : Juarez, Mexico

Peeron is incorrect this morning, but it looks like the http://tjum.anthill.de/geohash.html site is correct. Therefore, the Globalhash for August 18 is in Juarez, just across the border from El Paso Texas. This is the most urban globalhash I've seen. Anyone? --Thomcat 13:54, 17 August 2009 (UTC)

The sad thing is that it seems to be inside a house, i.e. of very questionable accessibility. Still worth a try if someone is within reach, I'd say. --Ekorren 14:05, 17 August 2009 (UTC)

2009-10-25: Firenze, Italy

The globalhash for Sunday is in an Italian field. Somebody go get it! --starbird 11:19, 24 October 2009 (UTC)

2010-01-12: Augusta, Georgia

The globalhash for Tuesday is on an Augusta, Georgia access road or joint driveway. This is EXTREMELY reachable. We need to find a local hasher! --Bill^2 15:59, 11 January 2010 (UTC)

LuxMundi & NamGeunGen are going to try to be there around 18:30. Also a couple of drag-alongs.

Globalhash Should Give a Week's notice: Discuss

A few people have proven that the global hash is not impossible, given patience, luck, and motorized transport, but as NWoodruff & LuxMundi, the first successful globalhashers pointed out, the time constraints forced them to use a car. We wouldn't have been able to make it to the Idaho globalhash at all if it hadn't had the extra lead time of being on a Sunday, and as it was we needed two different motorized vehicles to get there and back. Given a week's notice, we could have mounted a far more epic expedition, possibly including Rhonda, Xore, pi, JimL, Thomcat, Michael5000 and Chariot, mounting a geoevent atop that mountain like the world has yet to see. Given a week, I could have biked there from Vancouver!

Just as daily notice forces spontaneity in the regular geohashes, one week's notice forces spontaneity for longer trips. I would like more people to have the fun of a globalhash by making them easier to get to. Who has opinions on this idea? =Robyn 23:46, 29 March 2010 (UTC)

Geohashes are valid for one day. What if globalhashes were valid for one (ISO 8601) week and calculated using every Monday's Dow Jones and 2010-W13 etc.? --ilpadre 05:34, 30 March 2010 (UTC)
That makes seven times less globalhashes. I doubt, there would be more global expeditions. This weekend's globalhash wouldn't have happened, if only the Monday Dow Jones had been taken into consideration. I think, there still should be one global hash each day, but a time shift between calculation and actual happening is worth considering.--Reinhard 08:00, 30 March 2010 (UTC)

Exactly. My first idea was as Ilpadre suggested, but Reinhard is right: we want to roll the dice as often as possible in the hope of getting attainable coordinates once in a while. The Cancun coordinates were attainable last week,for example. Wade and I were planning to win the lottery and get first class tickets to Cancun for everyone, but with so little lead time it was impossible. -Robyn 17:57, 30 March 2010 (UTC)

I'd say the simplest thing would be to just extend the time: Each Dow opening one or more new globalhashes are generated, and you can go there anytime in the 7 days starting with the date used to generate it( as opposed to 24 hours currently.) I'd suggest we try that for a while. If it seems to have become too "easy", then we could tighten it down and say you have to go in a 24 hour period, but make it a week after the generating day to allow planning. Jiml 18:13, 30 March 2010 (UTC)

There's a hazard here: a large precentage of geohashes are at sea or in extreme northern or southern latitudes. By reducing the number of hashes, it might become extremely rare to get an even slightly makeable hash. You could allow points to linger for a week, I suppose. -- Wmcduff 18:22, 30 March 2010 (UTC)

That's why the proposal is not to reduce the number of globalhashes. A new one would be generated every day, just as now, but instead of 24 hours to reach it, you would have seven days. On any one day, you would have a choice of seven possible global hashes to attend: one brand new, one seven days old, and five in between. -Robyn 20:59, 30 March 2010 (UTC)

I'm currently not at home, so I'll postpone some thoughts for later. However, I want to file a far reaching opposition against those propositions. The fun in globalhashes is that they are so difficult and not possible to plan. Currently accessible globalhashes yield some significant excitement in the community: People try to find nearby hashers, who might be able to go there in time, and try to convince them to go, everyone cheers and follows the plans and reports as they come in. If we have a weeks notice, this will be all gone. Every theoretically accessible globalhash gets accessible for everyone. This will probably cause a few people spending ridiculous amounts on a few globalhashes, and everyone else feeling left out and losing interest in the whole thing. There will be nothing that special about the globalhashes any more. The globalhash is an ultimate challenge not every hasher ever gets a chance to do, and by that it's an ultimate achievement to get one. I can't see any reason why the rules should be changed in order to make this artifically easier (for rich people, that is, since the poor won't be able to go to the other end of the globe on a weeks notice either). Myself, I'm not going to take part in a game that is basically about being able to spend ridiculous amounts of money. - Ekorren appr. 21:20, 30 March 2010 (UTC)

Thank you for writing down your thoughts. They are quite equal to the ones I had after I wrote that it would be worth thinking about a longer preparation time. But it's not. I could repeat everything you wrote. It loses its specialty, its demands on quickly deciding and being spontaneous and – the worst thing about it – it would make those who can afford it travel ridiculous distances. Sure, I could take a flight to Spain for the weekend, if there's a globalhash announced some days before, but is this really what the globalhash was invented for? Thanks again to Ekorren. --Reinhard 21:39, 30 March 2010 (UTC)
I second Ekorren's sentiment but think that extends beyond financial resources to time resources as well. More so, I think it's okay for coordinates to be unobtainable, be they graticule or global coordinates, and not just for reasons of private property or conditions but also for reasons of time. Overcoming these challenges is part of what makes every successful geohash something to celebrate (and even some of the unsuccessful but epic ones too). - Trane 21:45, 30 March 2010 (UTC)
That's a good perspective, Ekorren. I was expecting the purist objection, but not those pragmatic reason. People shouldn't be able to buy a globalhash. I thought it was more likely that people could arrange to carpool, get cheap tickets or travel by bicycle if there were a longer lead time. After all, the more you have to spend, the faster you can get there, so give more time and you an get there more cheaply. As an example of what I was thinking, had we had a longer lead time for Sunday's hash, we could have brought along a third geohasher up to 200 lbs in weight in the helicopter at no additional cost. And we would have found out about it during the week, so that Rhonda could have done her weekend chores during the week, and then we could have all gone down Friday night, maybe in convoy with the Portland folks, hiked up, and actually camped on the spot for a sunrise hash. You're right that the change would allow the fabulously wealthy to travel to the otherwise unobtainable Antarctic geohashes, but I think that the real result would be having more people attend the reachable geohashes, or having the ones who do reach them be more awesome. There is already an inequity in that people with more means or more leisure time can travel to globalhashes more quickly. Does increasing the time really increase the inequity? -Robyn 21:57, 30 March 2010 (UTC)
Yes, it totally does. And I'm still not talking of hiring a plane here but of buying a plane ticket halfway through the world or a barrel of gas. Generally the challenge of a globalhash is not getting there, but getting there in time. Giving a weeks notice makes _every_ globalhash theoretically accessible. If you want enough planning time, do a retro globalhash. No one will seriously cheer for you or try to talk you into doing that, though, and this is for a reason. Ekorren appr. 22:08 UTC

While I concede that theoretically humans can reach any point on the Earth, most of them are laughable write-offs, too far out in the ocean or in the Antarctic. I would have said that the primary challenge of getting to a globalhash was in waiting for one to occur in a place that you could conceivably reach at all without a team from National Geographic and on a day that you and enough people to have a proper expedition party are available. I just wanted more people to have the opportunity to get to one when they do occur. However the vehemence of your opposition is greater than my zeal in advocating this change, so we'll just leave this discussion here, to dissuade the next person who might come up with the same idea. Thank you for defending the integrity of the globalhash. -Robyn 23:46, 30 March 2010 (UTC)

I also kind of like the idea of globalhashes being something almost unattainable. To me, like with Couch Potatoes, it is something where the algorithm has to select you, not the other way around. -- relet 07:45, 31 March 2010 (UTC)

Standard filenames and templates

I'm working on adding Globalhash support to Geohash Droid (yes, just in case you're crazy enough to try), and I'm wondering if we've agreed on filename standards (i.e. uploaded images) and a template to put in the wiki pages (like the current Expedition template). Do we have any standards like that yet?

If we don't yet, given the current naming scheme looks like "YYYY-MM-DD global" for the pages, I'd like to suggest the same as a prefix for files (so a file nominally called "Beans.jpg" would be uploaded as "YYYY-MM-DD_global_Beans.jpg"). Though I'm not sure what to use for the template yet. CaptainSpam 22:38, 30 May 2010 (UTC)

There isn't actually a prescribed scheme for images at all. It is expected that they contain the date and coordinates somewhere to make them unique enough, but the formatting varies. Anything including the date and 'global' should be good. The expedition reports so far took the form "YYYY-MM-DD global" and that's a good thing to keep. -- relet 22:41, 30 May 2010 (UTC)
All right, that sounds good, but have we made a variant of the Expedition template for globalhashes yet? Or should I just not put in a template for the time being? CaptainSpam 22:47, 30 May 2010 (UTC)
Hmm. The problem is that the map tag is a software module, not a wiki thing. So we mere wiki users cannot adapt that part of the template. Just add the coordinates and a link to some maps maybe. -- relet 22:51, 30 May 2010 (UTC)
Just a belated note to say that there is now the {{meetup global|date=DATE}} template, which gives the usual map of the (global) hashpoint and links, etc. — Benjw  {talk} 18:21, 4 September 2012 (EDT)

Better Distribution of Points?

Maybe the current algorithm is fixed in our minds now and there's too much inertia, but I've always been bugged by the uneven distribution of the globalhash points. True, this happens in the regular geohashes too, but within a particular graticule, the effect is minimal. Over the whole planet, we end up with the the points being too densely distributed near the poles and too sparsely distributed near the equator.

And the solution is pretty simple. Given the two decimal values (call them x and y) generated by the main algorithm, generate the longitude the same way: 360x-180 and adjust the latitude to even the distribution: arcsin(2y-1) (depending on whether you're using a degree-based or radian-based arcsin function, you might need to convert back to degrees.

I make this suggestion not because it will make it slightly easier to obtain a globalhash, but out of a perverse sense of purity.

ErWenn 16:13, 14 September 2010 (UTC)

I think your point is valid. I'm not sure I would support the proposal, because I don't think it matters in the end. I'm no great fan of the ugly patches we have in our algorithms so far (i.e. that the algorithm requires us to have a W30 rule in the first place), and I'll yet have to decide if yours makes things better or worse. I'd rather find a clean solution for the entire algorithm. -- relet 21:47, 14 September 2010 (UTC)
And it gives an incentive to all the researchers in the Arctic and Antarctica to go outside and grab themselves a rare and sought-after globalhash.Sourcerer 12:09, 17 May 2011 (UTC)
Man, I was thinking the same thing all day today at work... just discovered the globalhash, and this issue was bugging me. I thought there was a good chance someone had noted this earlier and posted it here, and sure enough I check here and ErWenn has! So I wholeheartedly second you ErWenn. Using latitude instead of sin(latitude) to choose the points is just plain wrong when you're on the whole globe. It's very slightly wrong for individual graticules where people live, but that doesn't bother me so much. (I'm an atmosphere scientist, not a mathematician.) --OtherJack 05:50, 27 January 2012 (UTC)
Hmm. I kind of like the idea, actually. -- relet 03:52, 28 February 2012 (EST)
I agree that arcsin should be used to make the globalhash distribution approximately uniform over the surface of the earth. This would only affect how the hash value gets mapped to latitude for the global hash - a function that is already unique. The only ugly part is that the original proposal for a globalhash didn't solve the distribution problem in the first place. --Robartsd 21:11, 3 February 2013 (EST)
I found geohashing a week ago, and this uneven distribution of globalhash has been bugging me since. Any ideas to convert to this arcsin version, perhaps provide both versions over a transition period? I live in Finland, so I'm not in the worst place regarding this, graticules are less than 30nm wide around here. Laurint 05:39, 12 September 2014 (UTC)
I certainly wouldn't be opposed! Though it has been the original way for many years now, so I'm also OK with keeping it as the original way for continuity's sake... Maybe we should make a proposal on the front page. -- OtherJack (talk) 10:51, 12 September 2014 (EDT)
No, no, and no again. Also, I oppose.
Globalhashes are good as they are: Easy to understand, a simple projection of the standard geohashing algorithm on all of the globe. Which was the idea. Just like geohashing. There's no "fairness" in the size of the standard graticules either, btw..
Also, a "perfect distribution" wouldn't be much "better". The average size of a graticule is 7870 km². The size of a graticule in central Europe is only about one tenth more than that. That means, in a "perfect" distribution by area, the globalhash would hit my graticule about once every 160 years instead of about once every 175 years. Even the largest graticules are just about 1.5 the size of the average graticules. So a "perfect" distribution would have hardly any effect on the reachability while changing everything away from the original idea and easiness. So why should we change it other than change for the sake of change? Why should we go away from the original idea?
Did I already say "no" to the idea? Well, if I did, I'll do it again.
And if you ask why I didn't say "no" earlier: Well, I never took that idea serious, and until now I was sure that it would die the death of ignorance like a huge number of other ideas that sprung from the minds of people who couldn't accept the challenge of geohashing and tried to propose something different yet easier. --Ekorren (talk) 11:18, 12 September 2014 (EDT)
OK, I understand your opposition... my problem was not with reachability (which of course is very difficult anyway, as you say!) but just with mathematical/aesthetic correctness on the sphere. Sorry I didn't make that clear. Anyway I'm not hugely opinionated about this, just thought it would be a nice idea... so unless Laurint or ErWenn feels as strongly as you we should probably just keep it as is. -- OtherJack (talk) 11:29, 12 September 2014 (EDT)
And, I think Laurint was also not so concerned about easiness or reachability... as he noted, he lives in Finland where graticules are small, so this change would make it harder for him to reach a globalhash, not easier... -- OtherJack (talk) 11:31, 12 September 2014 (EDT)
Dear OtherJack, I must disagree with your conclusion about easiness of reaching globalhashes: the algorithm produces uniform distribution of both longitude and latitude coordinates, i.e. they seem to be evenly spread out on equirectangular projection of world map [1]. But the closer you are to the poles, the larger portion of that map a circle with constant radius will cover, which means that there will (in the long run) be more globalhashes. Since it seems that I've been pondering this for almost 2.5 years, I am pretty sure I'm correct with this one Laurint, 12:37, January 20th 2017 (UTC)
Globalhashes are evenly distributed by graticule. But graticules are thinner towards the poles, so globalhashes are not evenly distributed by surface area. So what? Neither are ordinary geohashes, for the same reason. Yes, it could be fixed by making the algorithm much more complicated, but since it wouldn't actually make very much difference, simplicity should probably win out. Add into the mix the fact that hashpoints are often more difficult to reach in the higher latitudes because of mountains, wilderness, snow and stuff, those living in those places probably deserve more hashpoints near them.  :-) — Benjw  {talk} 15:11, 20 January 2017 (UTC)
Oppose: As it gets colder near the poles, there are more hashpoints, but harder to reach because of the temperature. It all balances out. --Sourcerer (talk) 20:16, 20 January 2017 (UTC)
Agreed with Benjw and Sourcerer... hahahaha. Also, I think you may have misunderstood me Laurint - of course I understand there will be more globalhashes within an x-km radius of you if you live near the poles than if you live near the equator. What I was trying to say back in 2014 was that you, Laurint, live quite near the Pole yourself - so the current arrangement actually benefits you! If we were to switch to arcsin, your globalhash frequency would decrease (even as mine, down here at latitude 38N, would increase.) So I was trying to defend you from the implication that you were trying to make your globalhashing job easier. :D -- OtherJack (talk) 21:18, 20 January 2017 (UTC)
(Btw it's actually a fun math problem to figure out the latitude separating the places that would "win" from those that would "lose" globalhashes due to the change... turns out it's not 45, but somewhere in the low 50s, around London or maybe Sourcerer's location. In the current system the chance of a globalhash landing in your local y km N-S by x km E-W box is basically y/(pole-to-pole distance) * x/(circumference of your latitude circle) = y/(piR)*x/(2piRcos(lat)) = xy/(2pi^2 R^2 cos(lat)), where R is earth's radius. Whereas in the future proposed system, it is the same everywhere, xy/(area of Earth) = xy/(4pi R^2). So if the rules were to change, one's chances would improve by a factor of (2pi^2 R^2 cos(lat))/(4pi R^2) = (pi/2) cos(lat). E.g. at latitude 0, your frequency would be multiplied by (pi/2) * 1 = 1.57, so 57% more "likely" to find a global point than you were before the rule change. But at Laurint's latitude of 60, it would be multiplied by (pi/2) * (1/2) = 0.79, so 21% less likely than before the change. The crossover is where (pi/2) cos(lat) = 1.00 exactly, so we need cos(lat) = 2/pi. The google tells me arccos(2/pi) is about 50.4597763 degrees, which lines up with the English Channel or central Germany.) -- OtherJack (talk) 21:18, 20 January 2017 (UTC)

A name for 'Globalocal' points?

Long explanation short, and cutting to the chase in this re-edit, I've been using "Globalocal" (or a variant of that) in my mind for "The point on the ground which, if it happened to be given as the local hashpoint within my graticule, would exactly be the globalhash point as well". And, not to be too insular about it, each graticule (with the same W30/DJIA-adjusted nature as the global calculation) has such a point. (Ok, so that excludes over 40% of the globe, already, but you get what I mean. ;)

(I'd also been playing with something like "gratiglobal" or "locicule", for the subset of the graticule area which represents how the graticule's own boundaries map onto the global algorithm's equivalent local coordinates. IYSWIM. But they're much less worthy neologisms.)

I finally put the above term down in writing, when I documented a 'near' miss in one of the upcoming dates for my graticule, where the global was 'only' a couple of graticules north and not far off eastwards (albeit in the midst of the North Sea), but surely it's an idea already conceived. So I was wondering if anybody has a (probably better) name for it? Is there a page link that I've missed? Can something like Peeron generate the points, or indicate the distances between global and local 'landing zones' automatically? I have all these questions and more, but I was intending this question to be briefer than in my prior edit, so I'll leave it there as a still open query. --Monty 14:26, 5 April 2012 (EDT)

We've played with the calculations (well, I have), but I haven't heard of any good name yet. How about something more dramatic. Like the "navel" or the "focal point" of the graticule? :D -- relet 14:31, 5 April 2012 (EDT)
Cheers, nice to know I'm not a lone madman. One way or another. Other terms I can remembering thinking of included: the "crux" of the graticule; the "homepoint"; or describing the point as a "hyperlocal", "metalocal" or "metahash" spot. But few of these satistfy me. "Globalocal" doesn't roll off the tongue so well, though. Or rather too well, but in danger of getting garbled in the attempt. ;) --Monty 15:48, 5 April 2012 (EDT)
This reminds me of the contraction mapping theorem; as such, perhaps these should be called fixed points? -- Phyzome 22:52, 3 February 2013 (EST)


(PS, as easy to tell you here as anywhere else, as you're doubtless about to respond. The Geco link complains that the date format isn't "%Y-%m-%d" format, but as far as I can tell the URL specifies the date that way.)

Which URL? The %m and %d formats are specified to have two digits at all times - that's a common mistake I can think of right now. -- relet 16:36, 5 April 2012 (EDT)
Straight from the Wiki's graticule page's templated link from below the map, I get sent to "http://relet.net/geco/?date=2012-04-05&lat=53&lon=-1&zoom=8" (have tested on another graticules, same result). And the full error message is "time data 'NaN-NaN-NaN' does not match format '%Y-%m-%d'" NotANumber? Unless it's not been string2int-ishly converted, for some reason, I can't work out why "2012-04-05" could be consider NaN. Sorry, off-topic. (Save that I was interested in looking at Geco's Globalhash marking capabilities, which I had only become aware of when I'd been reading some of the above Talk prior to posting. When this is resolved, shall we edit this diversion out again?)
Hmm, that link works for me. It might be something with the time zones... :| I'm going to try to confirm that. -- relet 17:59, 5 April 2012 (EDT)
For what it's worth, although I suspect that UTC reality should always have been honoured, together with whatever offsets are involved, I've just tried a switched from GMT+0 (or rather +1, i.e. British Summer Time/DST) to a US timezone (Mountain Time, randomly chosen). Also tried accessing from a neighbouring computer, another browser, of graticules in other bits of the world, with a date having (i.e needing) no leading zeroes in month and day. (Not all at the same time, though some piled up together in combination.) No other result, though. Already made sure I had a screenshot fragment (messagebox and URL behind it) but it contains no more information than I've already given you, apart from the error-dialog's title of "Message from webpage" maybe indicating a bit more about which subsystem (I'm guessing the .cgi (PHP?) interpreter) is complaining. Offline for an unknown length of time, shortly, so that's all I can do for you tonight. Sounds confusing. Hope it's not me and me alone equipped to open up an unnecessary can of worms. --Monty 18:45, 5 April 2012 (EDT)

2014-03-22: Waterloo, Ontario

2014-03-22 global is a Saturday globalhash that appears to lie in an agricultural area of Ontario less than 100 km from Toronto and about 120 miles from Buffalo. Someone close-by could say whether or not it is accessible. Bravissimo594 (talk) 00:15, 22 March 2014 (EDT) - Waterloo, Ontario - Google - OSM