Discussion What is the most efficient way to crowdsource the 3D system coordinates

[RedWizzard]

I've stuck my javascript implementation up on Github (here) if anyone wants to take a look. There are three pages that all run clientside:

systems.html: Lists all the reference systems, and calculates distances from the selected system to all other systems. You can sort the table by clicking on the column headers.

checkall.html: Runs my trilateration implementation for all the systems Michael provided using a set of reference systems. If you edit the file you can change which systems are used as references and the precision of the distances supplied to the algorithm. With only 2 decimal places and the right 6 reference systems it'll calculate the rest of the systems correctly.

entry.html: A mockup of what a data entry page might look like. It does the trilateration as soon as you change any data (once it has the minimum required 4 distances), which is really useful I think. If you enter a bad point you can immediately see which one is wrong because it will have the worst error value. Note this doesn't post any data back to a server - if someone wants to add that functionality go for it.

Get it here: zip file. Just unzip it somewhere and open the html files in a browser. Note: only tested with Firefox. Should work in Chrome or Safari but I wouldn't be surprised if IE chokes.

 

[wolverine2710]

I've stuck my javascript implementation up on Github (here) if anyone wants to take a look. There are three pages that all run clientside:

systems.html: Lists all the reference systems, and calculates distances from the selected system to all other systems. You can sort the table by clicking on the column headers.

checkall.html: Runs my trilateration implementation for all the systems Michael provided using a set of reference systems. If you edit the file you can change which systems are used as references and the precision of the distances supplied to the algorithm. With only 2 decimal places and the right 6 reference systems it'll calculate the rest of the systems correctly.

entry.html: A mockup of what a data entry page might look like. It does the trilateration as soon as you change any data (once it has the minimum required 4 distances), which is really useful I think. If you enter a bad point you can immediately see which one is wrong because it will have the worst error value. Note this doesn't post any data back to a server - if someone wants to add that functionality go for it.

Get it here: zip file. Just unzip it somewhere and open the html files in a browser. Note: only tested with Firefox. Should work in Chrome or Safari but I wouldn't be surprised if IE chokes.

SWEET. Nice javascript programming. Very nice indeed. Gonna play with it a bit. Works flawless it seems on FF 32.0.3. Laymen here. Question:
you wrote: "Calculated using 2 decimal places for distances". Why is that?
I thought on the galaxy map we have three decimals?
Was surprised with the mockup entry page. I had imagined an inputfield for 'current system' and then the 9 reference system. Any reason for doing it this way?

 

[RedWizzard]

SWEET. Nice javascript programming. Very nice indeed. Gonna play with it a bit. Works flawless it seems on FF 32.0.3. Laymen here. Question:
you wrote: "Calculated using 2 decimal places for distances". Why is that?
I thought on the galaxy map we have three decimals?
Was surprised with the mockup entry page. I had imagined an inputfield for 'current system' and then the 9 reference system. Any reason for doing it this way?

The 2 decimals thing was just to see how inaccurate the data could be, and how many reference systems the algorithm needed. With distances with 3 decimals such as from the galaxy map you can get away with fewer reference systems.

Since the entry page doesn't send the data anywhere I didn't bother with a field for the system name. You'd definitely want that in a real version. I'm planning to allow other reference stars to be used: there will be extra rows with a autocomplete dropdown for the star name. But having a bunch of standard references already to go saves time for the user.

 

[wolverine2710]

The 2 decimals thing was just to see how inaccurate the data could be, and how many reference systems the algorithm needed. With distances with 3 decimals such as from the galaxy map you can get away with fewer reference systems.

Since the entry page doesn't send the data anywhere I didn't bother with a field for the system name. You'd definitely want that in a real version. I'm planning to allow other reference stars to be used: there will be extra rows with a autocomplete dropdown for the star name. But having a bunch of standard references already to go saves time for the user.

Played around a bit with checkall.html and changed 2 into 3 decimals. Deselected a few sytems, selected 5 ref points (I believe) and then changed those (still having 5 reference points) and saw a change in errors. Gonna compare it to spreadsheet provided by tunamage and to his reference points. Also gonna see if it gets into problems when using the data used by Harbinger and for certain permutatios (post #206).

Looking forward to v2 where using a drop down box we can easily select another reference point (overrideng the defaults one in the js code) or add a new one. Might I suggest making those reference switchable (on/off). Easier then edit the js file. Also played around a bit with the entry.html and changed a value like x.969 into x.696 and saw the ref points change and also the errors. Another suggestion: its a bit hard to easily see which input field results in the highest error - perhaps some kind of color coding for the worst one or the last three worst ones.

So far Sir, I'm IMPRESSED.

I'm not sure how the accuracy is when compared with the other solutions (no code) but darn it looks SWEET and professional. Perhaps other authors can do something similar like your checkall.html, so that accuracy etc can be compared.

At some point we (community) have to decide what code or multiple code implementations will be used to calc the coords but your js version seems to be easily implementable in a web based version for the real crowd sourcing effort. Javascript is pretty fast - on my system.... checkall.html took about 115 msec.

Edit: What I like about the entry.html. One can directly see if for say 4 entries the three ref points are zero. I assume at that point the volunteer does not have to enter more values or? If the volunteer can't get 0 erros for the ref points he can change or add a ref point to see if then all is zero.
What I don't understand. The entry.html gives 3 ref points. I thought four were needed?

Edit2: Wanted to see how your solution compares with the that of Tunamage. Checked entry.html but can't find out how to change the reference points. it would'nt be as simple as changing sol into eranin in the displayed table would it?

 

[Mossgarth]

Thank you Tunamage. This I can relate to and adapt for my own system.

 

[wolverine2710]

Compared your results with results of Harbinger - as he has tested his algorithm with my dziban distances. He is using a different approach and he reported problems with some permutations. Tried two of them and no problems with your code. I tried another one, which gave result with the Harbinger code. In your case all three results are the same. In his case the coords fluctuate. BUT that was the whole point of his approach. Results:

Spoiler

This permutation has no result with Harbingers code. Please do not interpret the above as that the approach of Harbinger is not good. It sounds solid but the libraries used gave him problem iirc.

----------------------------------------------------
Distances
System Distance Calculated Error
Sol 74.486 0.000 (ref)
Flousop 66.846 0.000
Tring 77.256 0.000 (ref)
Rahu
Parcae
Hepa
Poqomathi
Hyperion
Aulin 46.908 0.000 (ref)
Calculated


Coordinates: (-62.15625, 38.46875, -14.3125)
RMS error: 0.000

Distances
System Distance Calculated Error
Sol 74.486 0.000 (ref)
Flousop 66.846 0.000
Tring
Rahu 33.315 0.000 (ref)
Parcae
Hepa 48.696 0.000 (ref)
Poqomathi
Hyperion
Aulin
Calculated

Coordinates: (-62.15625, 38.46875, -14.3125)
RMS error: 0.000

This permutation has no result with Harbingers code.
----------------------------------------------------
Distances
System Distance Calculated Error
Sol 74.486 0.000 (ref)
Flousop 66.846 0.000
Tring 77.256 0.000 (ref)
Rahu
Parcae 56.595 0.000 (ref)
Hepa
Poqomathi
Hyperion
Aulin
Calculated

Coordinates: (-62.15625, 38.46875, -14.3125)
RMS error: 0.000

All three results return exactly the same coords.
The code by harbinger did not. BUT that was expected and the way he tackles the problem.

Next: Check your results with the spreadsheet of Tunamage and then I'm done ;-)

 

[Athan]

Indeed... has anyone checked that working out the distance between pairs of MB provided stars results in the same distance between them as shown on the map ? If they all match, then the game would appear to be using the 1/32 ly grid, if not it's just MB's extract of the data that has that feature.

I've done this check for the first 20 systems in the file, measuring from Alderamin where I happened to be:

Name            Calcu Dist      Obs     Diff
Coquenchis       34.5455145876  34.546  -0.00048541
Mistana          78.9370360735  78.937  0.00003607
Gera             47.9949236411  47.995  -0.00007636
HIP 2422         81.6794061598  81.679  0.00040616
HR 8474          58.6434788974  58.643  0.00047890
Nyon T'ao Wujin  66.5959569672  66.596  -0.00004303
BD+71 1033       65.9160361985  65.916  0.00003620
Adepti           49.7557608945  49.756  -0.00023911
Kulkanabossongma 76.8641506572  76.864  0.00015066
HR 8423          84.2848418775  84.285  -0.00015812
LP 29-188        46.3741682035  46.374  0.00016820
Basium           52.5476067634  52.548  -0.00039324
Chirichianco     74.9157665002  74.916  -0.00023350
V740 Cassiopeiae 54.4522081164  54.452  0.00020812
Gliese 9843      82.2805853562  82.281  -0.00041464
HR 8133          90.4479520567  90.448  -0.00004794
Toyota           56.2744217818  56.274  0.00042178
Psamathe         34.6368718366  34.637  -0.00012816
Tapari           37.573963517   37.574  -0.00003648
35 Draconis      60.8345765427  60.835  -0.00042346

So the errors are on the order of 4*10^-4, or in binary ... in the 12th or so place after the binary point.

Correct me if you know better... but given I used sqrt((x1-x2)^2 + (y1-y2)^2 + (z1-z2)^2) to work out the distances, the error propagation of up to 1/64 on each input (the furthest a figure accurate to 1/32 could be in binary before it would switch to the next value anyway) means the error should be about 1/32 ?

This is pointing to the game actually only being down to 1/32 of a ly as well.

 

[JesusFreke]

To answer my own question, 3 digits distance holds across the galaxy. Puwaei NF-A D2 is 62149.547 LY from LHS 2764A (where I happen to be now)

If someone is in other places of the pill, it would be interesting if could post system and distance, and see if trilineration came back with something like -7763, 101, 61664, or if the angles become too small for it to work.

I tested this for a "fake" system that was at something like 30000, 40000, 40000 or something. I don't remember the exact details, but trilateration was still pretty accurate, although not accurate to the exact grid point any longer.

 

[Snuble]

Hey FD! We're nearly done with the milky way, can 3300 scientists please invent some way to travel to another galaxy (far far away)?

 

[TunaMage]

Correct me if you know better... but given I used sqrt((x1-x2)^2 + (y1-y2)^2 + (z1-z2)^2) to work out the distances, the error propagation of up to 1/64 on each input (the furthest a figure accurate to 1/32 could be in binary before it would switch to the next value anyway) means the error should be about 1/32 ?

I might also be incorrect here (and maybe also splitting hairs ),but I would say the max error on the equation could be bigger.

The max error would be (1/32)^2 per xyz coordinate set, so could accumulate to (3*(1/32)^2)^0.5 = +- 0.054 or about 1/18th.

I'm not sure how significant this is, and it is a boundary value.
...and like I said, I could also be talking rubbish Still need to dbl check.

 

[wolverine2710]

@Tunamage and RedWizard. Compared your results. Used the same inputs I used when I tested Redwizards results with Harbingers results. Current system is Dziban.

@Tunamage. Changes made

Spoiler

Had to change the downloaded as .xlsx version of your spreadsheet. Changed reference points so they are the same as the crhomatrix reference points. Added Sol to sheet 3. Had to change formatting for a few cells. The input fields set to three decimals. I'm a Dutchie and we use the , character as a decimal separator. So this caused everything to be round to 0 decimals. Changed calculated fields to 5 decimals so I could compare the result with the results of RedWizard.

Results:

Spoiler

Redwizard: Coordinates: (-62.15625, 38.46875, -14.3125)

TunaMage: Dziban -62,15570 38,47100 -14,31470
Rounded to 1/32 Grid -62,15625 38,46875 -14,31250
Diff: No difference between output RW and TM

Redwizard: Coordinates: (-62.15625, 38.46875, -14.3125)

TunaMage: Dziban -62,15480 38,47040 -14,31170
Rounded to 1/32 Grid -62,15625 38,46875 -14,31250
Diff: No difference between output RW and TM

Redwizard: Coordinates: (-62.15625, 38.46875, -14.3125)

TunaMage: Dziban -62,15750 38,46980 -14,31040
Rounded to 1/32 Grid -62,15625 38,46875 -14,31250
Diff: No difference between output RW and TM

Distances taken. Current system Dziban. Output copy/paste entry.html RedWizard.

Distances
System Distance Calculated Error
Sol 74.486 0.000 (ref)
Flousop 66.846 0.000
Tring 77.256 0.000 (ref)
Rahu
Parcae
Hepa
Poqomathi
Hyperion
Aulin 46.908 0.000 (ref)
Calculated


Coordinates: (-62.15625, 38.46875, -14.3125)
RMS error: 0.000

Distances
System Distance Calculated Error
Sol 74.486 0.000 (ref)
Flousop 66.846 0.000
Tring
Rahu 33.315 0.000 (ref)
Parcae
Hepa 48.696 0.000 (ref)
Poqomathi
Hyperion
Aulin
Calculated

Coordinates: (-62.15625, 38.46875, -14.3125)
RMS error: 0.000

Distances
System Distance Calculated Error
Sol 74.486 0.000 (ref)
Flousop 66.846 0.000
Tring 77.256 0.000 (ref)
Rahu
Parcae 56.595 0.000 (ref)
Hepa
Poqomathi
Hyperion
Aulin
Calculated

Coordinates: (-62.15625, 38.46875, -14.3125)
RMS error: 0.000

Results. Tested coords Dziban with 3 sets of different reference points. In all three RW cases result was the same. In all three TM cases 1/32 rounded results was the same. AND the results of RW and TM are the same.

Conclusion: Two very different approaches give exactly the same results - rounded to 1/32 LY.. Of course have only tested three cases but the checkall.html page of RW gives me a very good feeling ;-)

I'm done testing. Fingers bleed, need to stop. Very productive day it seems, wrt getting good results with the coords calcs.

 

[Athan]

I might also be incorrect here (and maybe also splitting hairs ),but I would say the max error on the equation could be bigger.

The max error would be (1/32)^2 per xyz coordinate set, so could accumulate to (3*(1/32)^2)^0.5 = +- 0.054 or about 1/18th.

I'm not sure how significant this is, and it is a boundary value.
...and like I said, I could also be talking rubbish Still need to dbl check.

Yeah, when I looked into errors estimation recently for something else the methods I found varied a lot, one seems to be what you've done (run the error values through the same equation they're used in). Although I note you used 1/32, when it should be 1/64 for the error. That makes it +/- 0.0270633....

But, still, the point is that if the supplied co-ordinates could be different from those used in-game by up to 1/64ly on any co-ordinate then we'd expect much bigger discrepancies between the calculated and in-game distances than I found.

We don't see that, therefore I conclude that in-game must also be using the same floating point format.

Also, I just checked wtbw's beta1 co-ordinates which he'd extracted using a debugger, and their fractional parts are also all multiples of 1/32.

So I think we're on safe ground assuming the in-game accuracy is to 1/32ly.

 

[Smacker]

I've updated my data to take into account Michael Brookes' data dump, and I also tweaked my algorithm to ensure that the final coordinates are aligned to a 1/32LY grid. Currently, I have trilaterated coordinates for 48 systems not included in Michael Brookes' data dump.

I have done a quick test using your coordinates in Trade Dangerous:

Local systems to LHS 457 within 11.0 ly.
----------------------------------------
ED   JF
2.94 2.94 CAER BRAN
3.34 3.34 CONNLA
5.05 5.05 CULANN
6.33 6.33 LHS 3281
6.66 6.66 WOLF 654
9.09 9.09 AUSTERN
9.60 9.41 LHS 3343
9.47 9.47 ROSS 868

Local systems to LHS 3343 within 11.0 ly.
-----------------------------------------
ED   JF 
7.28 7.18 LFT 1421
7.88 7.49 CULANN
8.32 8.32 NLTT 46621
8.38 8.38 LHS 3281
9.60 9.41 LHS 457

Local systems to CULANN within 11.0 ly.
---------------------------------------
ED   JF
 4.79  4.83 CAER BRAN
 5.06  5.05 LHS 457
 6.16  6.19 CONNLA
 7.88  7.49 LHS 3343
 9.48  9.42 LHS 3281
10.30 10.21 LFT 1421

So pretty close (only had time to do 3)
Worst error is about 0.4 Ly

 

[MadDogMurdock]

http://elitetradingtool.co.uk/

Thrudd has now added a new dialog adding new systems.

Input the distances to Sol and 3 other systems and it will work out the coordinates.

This uses a ranking system to access additional edit functions so the new system dialog may not be immediately available to new users.

 

[RedWizzard]

Looking forward to v2 where using a drop down box we can easily select another reference point (overrideng the defaults one in the js code) or add a new one. Might I suggest making those reference switchable (on/off). Easier then edit the js file.

I'll sort something out to allow selection of the reference systems. Probably checkboxes in the table.

Also played around a bit with the entry.html and changed a value like x.969 into x.696 and saw the ref points change and also the errors. Another suggestion: its a bit hard to easily see which input field results in the highest error - perhaps some kind of color coding for the worst one or the last three worst ones.

You can sort that table on the error column. That's what I was doing if I couldn't eye-ball it. It's mostly useful for catching typos. E.g say I'm doing Dahan but made a mistake on Flousop's distance (real distance is 31.680):
Sol: 46.324 calculated 46.330 error 0.006 (ref)
Flousop: 41.680 calculated 32.606 error 9.074
Tring: 119.181 calculated 116.234 error 2.947
Rahu: 31.825 calculated 31.818 error 0.007 (ref)
Parcae: 22.431 calculated 22.428 error 0.003 (ref)
I can see the error for Flousop is very high, so I can check and correct it. But it does need at least 5 distances: after I enter the first four the high error is actually on Sol (this is because Flousop and Sol are both in the same direction relative to Dahan). Essentially a high error means that the set of distances is inconsistent with the reference system coordinates.

Edit: What I like about the entry.html. One can directly see if for say 4 entries the three ref points are zero. I assume at that point the volunteer does not have to enter more values or? If the volunteer can't get 0 erros for the ref points he can change or add a ref point to see if then all is zero.
What I don't understand. The entry.html gives 3 ref points. I thought four were needed?

It's really the RMS error number at the bottom that you want to minimise. The three reference errors will always be pretty close to zero because these are the three distances that are doing most of the work to determine coordinates for the unknown system. Triliteration uses 3 distances to determine 2 candidate coordinates and then the 4th distance picks which candidate is correct. I'd recommend that users enters at least 5 distances with 3 decimal places: once you've got 4 distances entered the errors will be pretty close to zero (assuming no typos), if they stay close to zero with a 5th distance entered then you're almost certainly good.

My implementation has two extensions to the triliteration algorithm, both designed to take advantage of extra distances to increase accuracy. The first extension is that once it has figured out the two candidates from the three reference distances it doesn't just use one extra distance to pick between the candidates. Instead it calculates the error for each candidate over the entire set of distances. This reduces the chance of picking the wrong candidate (as in the case I was discussing with JesusFreke yesterday).

The second extension is that the trilateration code is run for all combinations of the 3 reference systems out of the entered distances rather than just picking the first 3. This means it's always using the best subset of reference systems it has data for.
E.g if you've entered distances to Sol, Flousop, Tring, Rahu, and Parcae, the algorithm will generate candidate coordinates based on each of these sets of 3 reference systems:
Sol, Flousop, Tring
Sol, Flousop, Rahu
Sol, Flousop, Parcae
Flousop, Tring, Rahu
Flousop, Tring, Parcae
Tring, Rahu, Parcae
Those 6 combinations will generate 12 candidate coordinates and then the best one is chosen based on the difference between the calculated distances and the supplied distances for all 5 systems. This is why you see the "(ref)" markers jump around as you enter data - an extra distance can mean that the best combination of reference systems changes.

Edit2: Wanted to see how your solution compares with the that of Tunamage. Checked entry.html but can't find out how to change the reference points. it would'nt be as simple as changing sol into eranin in the displayed table would it?

Yes, you can just edit the table at the bottom of the html file. The system names have to exactly match the data from Michael (you can get the exact name from the table in systems.html if you're unsure). It should just ignore systems it doesn't recognise though I haven't tested that

 

[Harbinger]

@wolverine2710: The reference stars you used seemed to cause the problems in the php code released by Snuble and modified by myself (perhaps due to being on the same plane or something?). I've been doing some testing using the form I set up here and in each attempt used all 9 point of reference stars that I chose myself due to their positioning in the capsule.

I've mapped the coordinates of 25 systems and I've not seen that error rear it's head in any attempt since. Here are the star systems I've mapped using this method:

• Perkwunos -40.500125, 46.749916, 6.749930.
• LP 274-8 -29.738246, 44.949649, 20.326609.
• Rho Coronae Borealis -29.874378, 42.095439, 22.217774.
• LP 386-49 -25.530545, 33.376693, 25.999697.
• LP 329-18 -26.624757, 39.874844, 23.280697.
• LHS 3124 -26.281543, 43.432673, 24.188797.
• LHS 399 -13.562117, 36.157422, 25.780629.
• LHS 396 -9.875744, 30.840733, 20.469069.
• OT Serpentis -11.125686, 30.341506, 18.407225.
• DE Bootis -7.437122, 32.625116, 16.999972.
• Veren's Stop -12.968789, 21.967734, 20.281170.
• G 181-6 -20.312616, 20.312786, 14.062621.
• Marcov's Point -22.342586, 14.626760, 17.655981.
• LP 229-17 -23.031609, 9.032729, 8.970076.
• G 203-51 -15.874793, 12.031296, 5.343640.
• Bidmere -7.436494, 13.219486, 1.904265.
• Wise 1405+5534 -8.030538, 13.438036, -1.844549.
• LHS 455 -16.905804, 10.219086, -3.437526.
• LP 71-165 -20.906919, 11.092810, -2.218044.
• LHS 465 -23.937852, 12.217715, -0.623736.
• Austern -25.155543, 15.344502, 9.374928.
• G 202-48 -15.531001, 14.375125, 1.875516.
• 2MASS 1503+2525 -6.187486, 17.999670, 8.124526.
• DG Canum Venaticorum -3.125089, 25.529649, 2.687266.
• LP 378-541 1.187879, 20.719592, 2.343413.

I don't think I've duplicated anything done previously.

My modified version of Snuble's php script here:
http://pastebin.com/q3Qf5uTX
(I named it trilat.php)

And the simple html form that passes data to it:
http://pastebin.com/BBjhW0Qh

This could easily be modified to store the inputs/results to a MySQL database etc.

 

[wolverine2710]

@RedWizard. Thanks for the answers, will look into them. A question though. I assume you didn't copy/paste all the 307 coordinates provided by Michael Brookes into your json structure, but used a small program for that. Could you share that one?

At this point I think we have to check the coords provided by the different commanders who spend time and love into this little crowd sourcing project. Your checkall.html seems to be perfect for that job. Afaik the most complete list of coords is to be found is in Trade Dangerous. Though I believe not all coords crowd sourced are in version 4.0.3 atm. It looks simpler to adjust your program then write my own. Though this might be THE time to learn python -as trade dangerous is written in Python.

 

[RedWizzard]

@RedWizard. Thanks for the answers, will look into them. A question though. I assume you didn't copy/paste all the 307 coordinates provided by Michael Brookes into your json structure, but used a small program for that. Could you share that one?

I just did it in Excel. Stick this formula into E2 and copy it down the column:

="{""name"":"""&A2&""", ""x"":"&B2*32&""", ""y"":"&C2*32&""", ""z"":"&D2*32&"},"

Then copy the contents of that E column into a text file and wrap it in "[" and "]". I decided to multiply all the coordinates by 32 so they were integers but that's not really necessary so you can take the "*32" bits out if you like.

I was impressed by Tunamage's algorithm so I've implemented that in javascript as well and modified the checkall page so you choose either his or mine when testing. Plus added the ability to pick the reference stars to use. His reference systems are a good choice and I can confirm that all the systems from Michael are correctly recalculated using that algorithm and those reference systems. Because he's only using 4 references any error in the data entry is impossible to catch (you just get bad coordinates), but that could easily be fixed by getting data for another system or two.

A new version of my stuff is available from the same location (here).

I'm going to make some changes to the entry.html page in the next few hours, then I'll start looking at what data is already out there and how it can be verified.

 

[wolverine2710]

@Harbinger. The reference points I used were from another commander - Chromatix who has put much time/effort in this thread. I lack the knowledge/time to determine if provided reference points by him or you are the optimal ones. I just use them.

That said, I changed the reference points.
Old : Sol,Flousop,Tring,Rahu,Parcae,Hepa,Poqomathi,Hyperion,Aulin
New (yours): Sol,Wolf 497,Huokang, Demeter, Clotti, Fu Haiting,San Guaralaru, Haras,Arabha

Iirc CM's names aside from being good were chosen so that they would easy to remember and type into the Galaxy Map.

I've entered your reference points into the checkall.html page of Red Wizard. In both cases ALL Michael Brookes coords result in a 0.00 error - had changed his 2 decimals into 3 decimals. So this is very nice to see.

Checked your newly created crowd sourced star systems with the ones in trade dangerous. Indeed they are all new. Must have taken you ages. Much appreciated. Will put your list in the trade dangerous thread so they can be put into TD by some volunteer.

Next stop: Just for fun. Going to insert your names/coords into RW's checkall.html. Very curious if they also come up with 0.000 errors. iirc your coords are also 1/32LY rounded, correct?

 

[wolverine2710]

I just did it in Excel. Stick this formula into E2 and copy it down the column:

="{""name"":"""&A2&""", ""x"":"&B2*32&""", ""y"":"&C2*32&""", ""z"":"&D2*32&"},"

Then copy the contents of that E column into a text file and wrap it in "[" and "]". I decided to multiply all the coordinates by 32 so they were integers but that's not really necessary so you can take the "*32" bits out if you like.

I was impressed by Tunamage's algorithm so I've implemented that in javascript as well and modified the checkall page so you choose either his or mine when testing. Plus added the ability to pick the reference stars to use. His reference systems are a good choice and I can confirm that all the systems from Michael are correctly recalculated using that algorithm and those reference systems. Because he's only using 4 references any error in the data entry is impossible to catch (you just get bad coordinates), but that could easily be fixed by getting data for another system or two.

A new version of my stuff is available from the same location (here).

I'm going to make some changes to the entry.html page in the next few hours, then I'll start looking at what data is already out there and how it can be verified.

Thanks for the update and answers. Also REALLY glad you are going to check all data which is currently out there. So I don't have to do that ;->. Harbinger posted 25 new names/coords ones and I believe yesterday another commander also had posted new names/coords here. I hope they end up soon in Trade Dangerous. You are probably are planning it but could you check the coordis in TD. This is the most complete list I found so far. Keep up the good work!!