Navigational Directions in space

[Ariphaos]

The fun part comes when you try to apply the same coordinate system to the edge of the galaxy compared to the center. As you get further out, you need more digits for distance and for location. Admittedly you are going to have more stars at the core of the galaxy, so it could be approved via majority vote, and to hell with what the Rim thinks.

No idea why this would be an issue. I have a hard time believing we would be stuck with double precision when quad precision is already available. This is only going to be used as a galactic coordinate system anyway - local systems will reference their star or the barycenter of their pair. This provides picometer resolution out to Andromeda.

[Oskuro]

From there you have a complex jump near the core to get around that darn singularity, then you get to coordinates that allow you a straight run to your destination.

Just a nitpick here. Since due to the rotation of the Galaxy, the core is somewhat flattened over the rotation plane, it might be interesting to circumnavigate it by going above or beyond it, rather than trying to fly around it on the galactic plane.


As for the precision issue, we're talking Galaxy here, an easy system to indicate Galaxy-wide coordinates will not have the precision for star-system coordinates, so as the order of magnitude of the distances involved gets lower, we would need to switch to local coordinate systems.

I don't think different coordinate systems are incompatible at all, nor do I think there's a need for a unified coordinate system for all orders of magnitude.

[Coalition]

The fun part comes when you try to apply the same coordinate system to the edge of the galaxy compared to the center. As you get further out, you need more digits for distance and for location. Admittedly you are going to have more stars at the core of the galaxy, so it could be approved via majority vote, and to hell with what the Rim thinks.

No idea why this would be an issue. I have a hard time believing we would be stuck with double precision when quad precision is already available. This is only going to be used as a galactic coordinate system anyway - local systems will reference their star or the barycenter of their pair. This provides picometer resolution out to Andromeda.

I am figuring that whatever precision you accept at 1000 ly will require 10 times the precision at 10* the distance. Alternately, you don't need as many digits for the precision at 100 ly. Figure the coordinates would be used to form 'exit volumes' for where interstellar travel would get converted over to system travel. So you would use system coordinates until you are outside the system volume (Earth might be 2* Neptune's orbit?), then you convert to galactic coordinates. You head to the target volume (using whatever method works) then convert back over to system coordinates.

Using angles the computer can look for items with similar initial digits as the ship's current position to see what could be in the way. I.e. if you are 57.4365 left and 2.8864 up, and there is a supergiant at 57.436 and 2.88, then the computer knows that it could be in the path. No math needed calculating angles, the computer just looks up in a database. It is primarily in the form of straight lies heading towards/away from the core, but the lower processing would be useful. You just need to make sure your nav database is up to date.

It does make the Core worlds very powerful trading centers, as a world that handles trade across a 10 ly cross section (as seen from the core) would cover a trading cross section 1000 ly at a distance 10* farther away. Lots of fun bribing the politicians to choose that method.

Lord Oskuro, you are right they can just jump over the galaxy then descend, I didn't think of that.

[Oskuro]

they can just jump over the galaxy then descend

I used to do that on Freelancer internet servers when some Clan decided to blockade a trade route and demand a fee to go through. Why go around through the asteroid fields and other hazards when I can go above or under them?
(Yeah, I did forego the faster trade lanes, but both map designers and players had a tendency to forget the Z axis)

[Beowulf]

Occurs to me that the military has already kinda solved a related problem: how to give the location of objects to arbitrarily high precision. Divide the world into map squares of approximately equal size, then coordinates inside the map square are given in the form xxxyyy, where the number of digits in x and y are equal. xxx gives longitude, yyy gives latitude. More digits increase precision. 10 digit gives accurate to meter precision. The map squares wrap around the world approximately in line with the latitude and longitude until you reach very high latitudes.

Something similar could be done with the galaxy, with map cubes, and locations inside the cubes being done in the form of xxxyyyzzz. xxx being distance side of the map cube in whatever units spinward, yyy being rimwards, and zzz being distance north. You can end up with arbitrarily good precision as necessary, and avoid having to get insanely high precision for the Rim. Map cubes could probably be labelled in the outer areas as a form of cylindrical polar coordinates (5 or 6 characters tells you distance from the center, angle around from the reference line, and height above the galactic plane). Map cubes would be conveniently sized so that for typical navigation you wouldn't normally have to leave the cube or the ones surrounding it.

[Oskuro]

Increasing precision in such a manner can be applied to mostly any system. In the model I proposed, where points are defined by a radius/angle pair, you can just change the units or use a bunch of decimals to achieve absolute precision, simplfying it with simple mathemathical conversions like x*10-15.

An interesting combination of a cube map system, and a radial system would be to define sphere sections to subdivide space, like making slices out of a watermelon. Of course that would be more useful in a fully spherical area, wich the galaxy is not.

[Beowulf]

Increasing precision in such a manner can be applied to mostly any system. In the model I proposed, where points are defined by a radius/angle pair, you can just change the units or use a bunch of decimals to achieve absolute precision, simplfying it with simple mathemathical conversions like x*10-15.

The problem is that for a radius/angle/angle triplet, you'd need to have much greater precision (and therefore number of digits) on the Rim, than at the Core. A map cube system requires only a constant number of digits no matter how far away you are from the Core.

[Junghalli]

Why would the need to take a detour to avoid the core be a major factor? The odds of a straight line from you to your destination going through the core singularity are pretty low. For the vast majority of possible destinations a detour would be totally unnecessary.

[Steel]

Increasing precision in such a manner can be applied to mostly any system. In the model I proposed, where points are defined by a radius/angle pair, you can just change the units or use a bunch of decimals to achieve absolute precision, simplfying it with simple mathemathical conversions like x*10-15.

The problem is that for a radius/angle/angle triplet, you'd need to have much greater precision (and therefore number of digits) on the Rim, than at the Core. A map cube system requires only a constant number of digits no matter how far away you are from the Core.

How is this any more of a problem (ie not a problem at all) than the fact that in cartesians you "need more digits" to represent (10,10) than (1,1)?

[Hawkwings]

The problem with setting up "subsectors" is that it requires standardization and cooperation from all the people that govern known space. Sure, this is easy if you're the Imperium of Man, but not all spacefaring cultures are so lucky. Having a single system centered around the core/seat of government allows all your ships to know exactly how to navigate, and ignore whatever navigation system the locals in subsector Nowhere-D have chosen to set up. because of this, I'd imagine that at least military ships would favor a centralized navigation system based on one set of reference points only.

An alternative would be to find easily-recognizable and noteworthy (ie large energy output) stellar phenomena, about 6-8 scattered around the galaxy, and use those as natural navigation aids. Set up a radius/angle/angle system for the galactic core and each one of these stellar phenomena. Then you can map things in relation to multiple reference points, as well as navigate using whichever reference point is closest. And your computers could translate galactic coordinates to Reference-Beta coordinates when you're near the beta reference point.

[Master of Cards]

For the people with resolution problems would it be easier to divide the galaxy into 10/0/0 lightyear cube sectors then have a sector level coordinate system? This would allow for more precision and would allow for better orangzation of planets by allowing multiple planets having similar names be in different sectors so "Helm set course for Jameson system: wouldn't be heard as "Helm set course for Thameson system"

[GrandMasterTerwynn]

Increasing precision in such a manner can be applied to mostly any system. In the model I proposed, where points are defined by a radius/angle pair, you can just change the units or use a bunch of decimals to achieve absolute precision, simplfying it with simple mathemathical conversions like x*10-15.

The problem is that for a radius/angle/angle triplet, you'd need to have much greater precision (and therefore number of digits) on the Rim, than at the Core. A map cube system requires only a constant number of digits no matter how far away you are from the Core.

How is this any more of a problem (ie not a problem at all) than the fact that in cartesians you "need more digits" to represent (10,10) than (1,1)?

If you went with radius/angle/angle, a milliradian angle at 1000 light-years from the Galactic center will get you 1 ly worth of resolution. At the edge of the Galaxy, 50,000 light-years from the core, each milliradian of angle change gets you 50 ly. To get the same 1 ly resolution at the Rim, you'd need to specify your angle down to the tens of microradians, which gives you a light-week of resolution 1000 light-years from the Galactic center. Dividing the galaxy into 'map cubes,' or 'sectors' in sci-fi speak seems to be a much cleaner solution. You could divide the galaxy into cubical sectors, each one 50 light-years on a side (I picked this value because a milliradian angle divides the galaxy into 50 ly arcs at the galactic rim.) Each one would be given a number of the format xxxxzzzzrrrr, where x is the angle from some arbitrary reference axis parallel to the galactic plane passing through the galactic center, z is an angle from the reference axis perpendicular to the galactic plane, passing through the galactic center; and r is the sector of interest along the line described by these two angles. So, in this example, if we were to assume that the x axis passes beneath Earth, Earth would be located in Sector 000000340520, since we are lined up with the x axis, are something like 900 light-years above the galactic plane, and we're 26,000 light-years from the galactic center.

Once you are within a given sector, navigation becomes relative to some central reference point within the sector. You could subdivide the sector into some arbitrary number of cubes. This would get you into a given starsystem, where in-system navigation methods would apply.

[Gil Hamilton]

The problem is that for a radius/angle/angle triplet, you'd need to have much greater precision (and therefore number of digits) on the Rim, than at the Core. A map cube system requires only a constant number of digits no matter how far away you are from the Core.

Are more digits really that big a problem for a society where one is really concerned about locations in the Core and in the Rim? Spherical coordinates are still more intuitive that a map cube when discussing motion in space.

[erik_t]

Gil, why are you insisting on spherical instead of cylindrical? Until we start wandering into the Magellanic clouds, I see no reason to disregard the obvious geometric preferences of the Milky Way.

[Isolder74]

Gil, why are you insisting on spherical instead of cylindrical? Until we start wandering into the Magellanic clouds, I see no reason to disregard the obvious geometric preferences of the Milky Way.

Well there are the globular clusters out there as well.

[GrandMasterTerwynn]

Gil, why are you insisting on spherical instead of cylindrical? Until we start wandering into the Magellanic clouds, I see no reason to disregard the obvious geometric preferences of the Milky Way.

Well there are the globular clusters out there as well.

As globulars are composed entirely of old Generation I stars, there is pretty much nothing of interest in them to an interstellar civilization requiring easy access to elements heavier than carbon.

[Steel]

How is this any more of a problem (ie not a problem at all) than the fact that in cartesians you "need more digits" to represent (10,10) than (1,1)?

If you went with radius/angle/angle, a milliradian angle at 1000 light-years from the Galactic center will get you 1 ly worth of resolution. At the edge of the Galaxy, 50,000 light-years from the core, each milliradian of angle change gets you 50 ly. To get the same 1 ly resolution at the Rim, you'd need to specify your angle down to the tens of microradians, which gives you a light-week of resolution 1000 light-years from the Galactic center.

Again, this does not actually show how a cylindrical polars system would be a problem.

All that happens is that you would choose however many digits were required to specify somewhere at the furthest distance you are interested in to the required precision (IN EXACTLY THE SAME WAY YOU WOULD DO FOR A CARTESIAN COORDINATE SYSTEM) and then send the thing as (000003,angle blah, z coord) for a core system and (283723, another angle, z coord) for somewhere on the outer rim. That was tricky wasnt it!

In a cartesian system the coordinates would look like (000003,000007,000002) for a core system and (123456,789123,456789) for an outer system anyway.

The cylindrical system has the virtue of being far nicer for the symmetry of the problem.

[Gil Hamilton]

Gil, why are you insisting on spherical instead of cylindrical? Until we start wandering into the Magellanic clouds, I see no reason to disregard the obvious geometric preferences of the Milky Way.

My preference was spherical instead of cartesian, actually. Cylindrical is less bad. However, my use of spherical coordinates for galactic coordinates is an extension of using the same system for solar systems, where spherical coordinates are the most intuitive (and, in fact, already used in astronomy).