Re:Postgres geography type is not limited to Earth
by askvictor on Sunday August 20, 2017 @08:36PM
(#15061094)
This is awesome. Purely out of interest, how would one go about a similar location database for arbitrary locations in space? PostGIS presumably assumes a spheroid planet; what if I'm tracking objects in interplanetary space? Are there any DB plugins (accessible outside of national space agencies) that can do that?
Re:Postgres geography type is not limited to Earth
by Demiurge on Monday August 21, 2017 @02:51AM
(#15062459)
That would require arbitrary telemetry and possibly non-deterministic solution, or something like SPICE(https://naif.jpl.nasa.gov/naif/toolkit.html) which is a bit too far outside of 'database' domain.
Proper motion is rather slow. Fortunately astronomy can take measurements across the centuries.
"Proper motion was suspected by early astronomers (according to Macrobius, AD 400) but a proof was not provided until 1718 by Edmund Halley, who noticed that Sirius, Arcturus and Aldebaran were over half a degree away from the positions charted by the ancient Greek astronomer Hipparchus roughly 1850 years earlier."
Re:Postgres geography type is not limited to Earth
by haskal on Monday August 21, 2017 @03:29PM
(#15067088)
For a person who is relying on benchmarks, you should know that benchmarks are only a rough indication of how well anything performs for a very specific use case.
Your reply is arrogant and ignorant, and you should feel bad about typing that.
> If you are that interested do your own tests.
Why is he not allowed to review your results or check what factors you benchmarked it on?
If there's no citation, that claim does not stand.
Re:Postgres geography type is not limited to Earth
by chillydawg on Monday August 21, 2017 @08:07AM
(#15063923)
To be fair, Oracle is usually extremely good at most of the stuff it does. I've no idea about GIS specifically, but, generally, you get what you pay for.
Re:Postgres geography type is not limited to Earth
by adamnemecek on Sunday August 20, 2017 @07:26PM
(#15060770)
I'm being nitpicky but "geography" means "description of Earth" so if we aren't talking about Earth it's not geography anymore :-). The word might be topography?
Re:Postgres geography type is not limited to Earth
by falcor84 on Sunday August 20, 2017 @11:00PM
(#15061619)
Ha, I personally always assumed that "earth" in this context refers to land/soil. Could anyone familiar with the Greek root let me know if that works in the original?
As you can see, the greek word refers equally well to the planet (sense I, "the earth, including land and sea, in contrast to heaven") and the dirt (senses 2, III, and IV), just like the english word.
Re:Postgres geography type is not limited to Earth
by brennankreiman on Sunday August 20, 2017 @11:25PM
(#15061706)
Phil Stooke has an interesting method for modeling small space objects, would be interesting to apply a non-spherical coordinate system. http://publish.uwo.ca/~pjstooke/plancart.htm
I guess you'd just want to store the orbital information of that object, as it probably won't stay at a fixed position.
Sensibly, the star catalog is arranged by apparent position from Earth. There are two kinds of motion that need to be accounted for: the precession of the earth, which affects all coordinates uniformly ( http://www.stargazing.net/kepler/b1950.html ) and "proper motion" of the stars themselves: https://en.wikipedia.org/wiki/Proper_motion
Proper motion is rather slow. Fortunately astronomy can take measurements across the centuries.
"Proper motion was suspected by early astronomers (according to Macrobius, AD 400) but a proof was not provided until 1718 by Edmund Halley, who noticed that Sirius, Arcturus and Aldebaran were over half a degree away from the positions charted by the ancient Greek astronomer Hipparchus roughly 1850 years earlier."
For example, in Korea a bunch of mapping providers use "UTM-K", which is some custom UTM projection
(defined such that coordinate [1_000_000, 2_000_000] is some point in the mountains in the north of South Korea (38.0, 127.5 lat/lon))
Just search UTM-K on spatialreference, http://spatialreference.org/ref/sr-org/7308/
And you get the magic postgis incantation http://spatialreference.org/ref/sr-org/7308/postgis/
Then just set the srid and everything should work nicely.
Meanwhile, they both are missing the concept of spatial contexts.
If you are that interested do your own tests.
Your reply is arrogant and ignorant, and you should feel bad about typing that.
> If you are that interested do your own tests.
Why is he not allowed to review your results or check what factors you benchmarked it on? If there's no citation, that claim does not stand.
That's a fascinating/contrary, albeit nebulous thing to say in a thread about Postgres...
https://en.wiktionary.org/wiki/areography
> "The scientific study of the planet Mars' geographical features;"
Wait, theres that word again...
Also Topography refers to physical features whereas Geography is much more broad. Funny that there doesn't seem to be a single word for it.
"γῆ • (gê) f (genitive γῆς); first declension
1. land, earth 2. country 3.soil"
I. earth opp. to heaven, or land opp. to sea; κατὰ γῆν on land, by land
2. earth, as an element, opp. to air, water, fire
II. a land, country; γῆν πρὸ γῆς from land to land
III. the earth or ground as tilled
IV. a lump of earth, in the phrase γῆν καὶ ὕδωρ αἰτεῖν, γῆν καὶ ὕδωρ διδόναι, in token of submission
http://www.perseus.tufts.edu/hopper/text?doc=Perseus%3Atext%... is a much more detailed, less quotable dictionary entry.
As you can see, the greek word refers equally well to the planet (sense I, "the earth, including land and sea, in contrast to heaven") and the dirt (senses 2, III, and IV), just like the english word.