RV Sonnenkreis - Decoding Universal Cartographics

[Jackie Silver]

I think your 36 might be a 35... If you check sectors_0plus, row 26:

At the left end we have Glaitu, with 5 unknowns to the left, and then Blaicheae. At the right end we have Glaicheae followed by Braitu. That Gl run, of what we can see, is 35 long.
In between Bl and Gl should be Cry, which elsewhere you identified as being 5 long. That makes up those five unknowns, meaning the run actually is 35 long.

That does also fit very cleanly - for whatever reason the run starts at -tu and ends at -cheae. That is the case for Gl, looks like it may be the case for Bl before it, and could also be the case for Br after it.

That doesn't help us with all the crazy edge cases though.

Edit: I am still completely stumped on how continuations between stacks work (if they happen at all). It doesn't seem to cleanly carry the pattern on, but it also doesn't always seem to start a fresh pattern at the start of the new stack - although it sometimes does, so there's a chance there are just a lot more custom lengths we haven't nailed down yet.

Here's my efforts from today - I've been following lines along hoping to find the extents, or at least to constrain them. "rfpp mod" has the list of phonemes in order. (so far as I can tell, the order from the pgdata list is correct, although there's a number of places I haven't been able to follow definitely) The "length lower minimum" is the minimum length of the sequence in sectors; the "length upper maximum" is the maximum length possible if the sequence took up all adjacent sectors that are in the two-part-name format. There is a great confusion of lengths! You're probably right about the 35 limit.

At this rate we'll have most of the zero plane mapped before we get the full sequence... or worse.

(Imagine a loud and repeated string of cursing. My laptop is dying, and when LibreOffice crashes - which it's done twice now - it's killing not only the file I'm working on, but also any changes made since the file was opened (even if saved, which I reflexively do constantly) and the autorecover doesn't work. So I've lost several large strips of data, about a half hours worth, twice. Not amused. On the plus side I did manage to confirm that OU is of length 35 before losing that strip, so 35 is looking even more promising.)

 

[Jackie Silver]

I think I've got the boundaries for the planes - a 120 x 120 sector box. If we start at (0,0) then WREGOE is at (60,39) and putting STUEMEAE at (60,60). Quite a lot of open space at the ends. Working so far, though it's hard to find sectors whose lengths and positions are known exactly enough to check.

(edit) Scratch that, badly wrong here.

(edit) I'm an absolute muppet - I had my sector maps wrongly aligned with each other, which means there's a lot more space between them. A 154 x 154 sector box with WREGOE at (76,66) and STUEMEAE at (76,76) may be more like it, but I had to make some guesses as to the lengths of sectors - if J,SQ,K,L are 35, Pyth is 1, M,ST,N are 35 then that fits.

 

[Esvandiary]

(edit) I'm an absolute muppet - I had my sector maps wrongly aligned with each other, which means there's a lot more space between them. A 154 x 154 sector box with WREGOE at (76,66) and STUEMEAE at (76,76) may be more like it, but I had to make some guesses as to the lengths of sectors - if J,SQ,K,L are 35, Pyth is 1, M,ST,N are 35 then that fits.

154 is very close to 156, which is double the nominal width of the galaxy... No guarantee it's not slightly wider on one side though, I suppose!

 

[Jackie Silver]

Could very well be 156, there's a little leeway depending on where exactly the sequences start. I'm going to get some more data from the -1 plane which should help clear it up - unfortunately I'll need to repair my antique laptop first.

(Wrong again - but getting there!)

 

[Esvandiary]

Could very well be 156, there's a little leeway depending on where exactly the sequences start. I'm going to get some more data from the -1 plane which should help clear it up - unfortunately I'll need to repair my antique laptop first.

Best of luck... Beat it into submission!

In your last edit, was Wregoe meant to be at (76,55) rather than (76,66)? Otherwise I'm confused
Just want to make sure I understand how the 154x154 (or 156x156) grid fits.

 

[Jackie Silver]

Best of luck... Beat it into submission!

In your last edit, was Wregoe meant to be at (76,55) rather than (76,66)? Otherwise I'm confused
Just want to make sure I understand how the 154x154 (or 156x156) grid fits.

This is where I'm going with it - I've put the -1 and 0 plane maps side by side, with a lot of space between them, and I'm trying to narrow down the sector sizes and positions well enough that I can match up the eastern/right hand end of the -1 plane with the western/left hand end of the 0 plane. There's generally at least one full set of sectors between them as there's a lot of space, but there's probably too much space in the spreadsheet as it stands, and it won't be as large as 156 after all. Getting there slowly.

This is the latest set of sector lengths that I've got.

 

[Esvandiary]

AHHHHH REALISATION [wacky]

The assumption for class 2 sectors is that it acts as though it's done two full runs of the 64-long pattern; when it continues two stacks up, it's exactly 16 suffixes further along the list, or two full runs.

I was just looking at the continuations in your merged 0 and -1 list... I just tried two different runs and found that in both cases, it only made sense if beyond the 78 normal sectors, there was an extra gap of 50 between (X, -1, 38) and (X, 0, -39).

What's 64 x 2? 128.
What's 78 + 50? 128.

I think we have our width. I feel like an idiot for not thinking of this before
If it makes you feel any better, I literally just held my hands to my head and said "OH " quite loudly when I realised. Hopefully it didn't wake up my better half...

Edit 2: I believe the Z component is (in some fashion) encoded in phoneme 3... Towards the -Z end the third phoneme is near the start of that infix list, and towards +Z it travels towards the end of the list. I imagine it's just incremented whenever it runs out of options in a run. It also appears to checkerboard in Z slices usually 1-2 apart.

 

[Vitamin Arrr]

OK I'm almost done with my survey of M36 sector. I found something I've never seen before: a trio of suppressed D-named (?!) stars:

M36 SECTOR EB-X D1-32
A - TTS3Va, 3134k, 0.4336 solar masses, 0.6370 radius, 180 million years old

M36 SECTOR AV-Y D1-36
A - M1Va, 3481k, 0.5039 solar masses, 0.6152 radius, 1.824 billion years old

M36 SECTOR EB-X D1-29
L1V, 1917k, 0.1836 solar masses, 0.3240 radius, 1.28 billion years old

We've definitely seen suppressed F, G, H... but D? Have you seen that before Jackie?

..

I also developed an algorithm to determine the starting point (lower southwest corner) of a sector. Given a sufficient number of sector names and their precise coordinates in the galaxy, it reverse-calculates the average predicted "starting point" for each system, then averages all of those against each other to reach a refined "starting point" within a margin of error that shrinks the more data you put in. If one of you can get me a JSON dataset of procgen names and the XYZ coordinates of the systems, I can feed that into the algorithm and spit out a JSON dataset with the "starting point" of each sector. (Of course if you've already done that then just ignore me.)

..

Also I downloaded your latest spreadsheets and I'm going to see what I can contribute to helping you decode that. I pulled down some machine learning algorithms from github and I'll see if there's any pattern detection ones that could be of use. Going to go back through this thread and see if I can figure out what you guys have going on here.

Especially since your laptop is dying Jackie, and multiple folks are now collaborating on this project, have you thought about migrating to Google Sheets? That way if your laptop dies then you won't lose anything. It is also beneficial for collaboration. I have a friend who could also help, he is pretty much always exploring and seems to want to help with data gathering.

 

[Jackie Silver]

We've definitely seen suppressed F, G, H... but D? Have you seen that before Jackie?

I don't remember off-hand so if I have they must be rare - looking at the log for STUEMEAE the D stars listed are unsuppressed and I'd think they'd be suppressed there. I'll look at those M36 stars now.

I also developed an algorithm to determine the starting point (lower southwest corner) of a sector. Given a sufficient number of sector names and their precise coordinates in the galaxy, it reverse-calculates the average predicted "starting point" for each system, then averages all of those against each other to reach a refined "starting point" within a margin of error that shrinks the more data you put in. If one of you can get me a JSON dataset of procgen names and the XYZ coordinates of the systems, I can feed that into the algorithm and spit out a JSON dataset with the "starting point" of each sector. (Of course if you've already done that then just ignore me.)

Sounds interesting - one of Alot's posts above has got a thing that pulls the JSON system list from EDSM. (Sorry I lost my copy of it along with the rest of the stuff on my laptop or I'd link it! Haven't got round to reinstalling everything and getting it all back to the right place yet.) He's used the coordinates to fit everything there into the sector grid maps we're using. Something to work out the sector start position from the system names would be good for looking at the handplaced sectors though.

Especially since your laptop is dying Jackie, and multiple folks are now collaborating on this project, have you thought about migrating to Google Sheets? That way if your laptop dies then you won't lose anything. It is also beneficial for collaboration. I have a friend who could also help, he is pretty much always exploring and seems to want to help with data gathering.

Yes. I've half-sorted my laptop now at least. What I was thinking is that once we're absolutely certain of the dimensions of the sector map, I'll put together all six planes into one giant spreadsheet running left-to-right and then I'll do battle with Google docs; I'm a luddite when it comes to that sort of thing though.

Edit 2: I believe the Z component is (in some fashion) encoded in phoneme 3... Towards the -Z end the third phoneme is near the start of that infix list, and towards +Z it travels towards the end of the list. I imagine it's just incremented whenever it runs out of options in a run. It also appears to checkerboard in Z slices usually 1-2 apart.

That's just confusing - the logic behind the way they've done things is strange! I had a quick look at 128 yesterday and it looked quite "tight", but then I'm still working on getting the exact sector lengths. Will get back to that now. Certainly 128 x 128 sounds like an excellent round number! Do the checkered two-part names start neatly at one corner of a 128 x 128 grid? I couldn't work it out.

 

[Esvandiary]

I also developed an algorithm to determine the starting point (lower southwest corner) of a sector. Given a sufficient number of sector names and their precise coordinates in the galaxy, it reverse-calculates the average predicted "starting point" for each system, then averages all of those against each other to reach a refined "starting point" within a margin of error that shrinks the more data you put in. If one of you can get me a JSON dataset of procgen names and the XYZ coordinates of the systems, I can feed that into the algorithm and spit out a JSON dataset with the "starting point" of each sector. (Of course if you've already done that then just ignore me.)

If I understand correctly, that would be very useful for finding the coords of the "not fully PG" sectors such as Col 359 etc. I think we can already reliably find the origin of any fully PG sector from a single set of coords within it, since they're aligned to the 1280Ly boundaries (the get_sector function in pgnames.py in my tools does this).

That's just confusing - the logic behind the way they've done things is strange! I had a quick look at 128 yesterday and it looked quite "tight", but then I'm still working on getting the exact sector lengths. Will get back to that now. Certainly 128 x 128 sounds like an excellent round number! Do the checkered two-part names start neatly at one corner of a 128 x 128 grid? I couldn't work it out.

I've done some more "linking" between -1 and 0 assuming 35 being the full length, and so far there are about five runs which match perfectly, a few more which match within margin of error, and no counterexamples.
I agree that the logic for the naming here is really strange... The suffix (phoneme 4) changes most often, but then the one that seems to change most often after that is the prefix! Phoneme 2 seems to change least often, and phoneme 3 is in the middle - which is why it seemed to be a pretty good indicator for changing Z.
Edit: Not sure about the phonemes which change most often actually... We've got a pretty limited view from 0 and -1 (and I have an unfortunate feeling we actually have quite a limited view from -3 to +2...)

I'm already starting to see a bit of evidence of checkerboarding, however, so I think we're back on that ol' train of fun.

 

[Jackie Silver]

I haven't put all the maps together yet but I've looked at the approximate extent of star-populated sectors in the 0, 1 and 2 planes; the radii are approximately 30 sectors, 20 sectors and 10 sectors.

This gives us a total of roughly 2 * pi * (10**2 + 20**2 + 30**2) sectors, ~ 9000 star-populated sectors, so that's hopefully a reasonably good estimate for Marx's question above.

 

[Vitamin Arrr]

Interesting, Jackie.. almost 9000, eh. So, two thoughts:

1) If indeed they used some kind of "mass function" to seed the stellar forge, I find it interesting that they have appx. 115 phonemes. Because y'know what other popular list also has ~115 items in it and relates to masses? The Periodic Table of Elements. I wonder if perhaps there is some kind of weird correlation here?

2) I think there will end up being exactly 8,480 sectors if my current guess is right. Each one would have three bytes of data to represent its sector name. E.g. 0x7F 0x5D 0x2F. I could be way off.. gonna check more later.

 

[Jackie Silver]

Here is the latest version of the sector map: all six planes of sectors arranged side by side, with each sector in a 128 x 128 grid. Everything seems to fit up nicely so far.
(If you're wondering, sectors in lower case are ones trawled from EDSM, sectors in capitals are usually ones that I've added; sectors highlighted in yellow are ones that I've added recently.)

If this looks ok to everyone I'll get it onto Google Docs. Once I work out how to use Google Docs.

Also here is the latest phoneme table.

For the rest, I'm on course for TZAIWNS and moving fairly quickly; I've crossed back out of the wilderness to the edge of the spiral arm so my route plotting is becoming more linear and less like a frog at a disco.

Incidental music: the Stone Roses (on loop), the Smiths, Simon & Garfunkel (you can blame whoever was taking the mick out of the Queen of Sadge for that one...)

 

[Pirin]

Jackie, as promised here is my data on big and fat stars Sorry for the delay, had to search all my folders and compile this spreadsheet.
Stars with capital letters are the main stars in their system.
Attention: long table

Spoiler

System	Primary Star	Age	Mass	Size	Temperature	Other 1st Tier Stars	2nd Tier Stars	3rd Tier Stars	Note
BLO AIP AA-A H22	B2 IVA	<1	87.8008	287.3694	27269	O9 IIIA, O7 IVA
CLEEQUE AA-A H4	WO0 I	<1	112.8398	4.8282	40815
CLEEQUE AA-A H64	WO0 I	<1	85.8789	4.0491	61249	O0 VZ, O6 IVA
CLEEQUE AA-A H313	O9 IIIA	<1	94.9648	208.2875	34728
CLEEQUE AA-A H607	O9 IIIA	<1	62.9609	203.2827	35191
DRYI AUDST AA-A H103	WNC0 I	<1	104.3242	4.7526	1559				Red colour
DUMBAI AA-A H26	WNC0 I	<1	75.7188	8.4934	49024	WO0 I
DUMBAI AA-A H51	WO0 I	<1	67.5469	7.0491	42047	B6 IVA, B0 VZ
dumbai aa-a h51	B6 IVA	<1	31.113	393.4711	19073	WO0 I, B0 VZ
ELLAIRB AA-A H5	AEBE3 VB	2	49.1914	0.5458	4640		TTS8 VI, TTS7 VI, TTS0 VIx2, TTS6 VIx2, TTS1 VIx2, TTS9 VI
EOS PHYLOEA LW-M D7-967	S5 IIIB	12878	0.9570	29.4511	2803
HD 164662	K1 III	6184	0.7109	19.7293	4,264.0	M2 VA
HYPOAE AUDST AA-A H189	O9 IIIA	<1	76.1563	203.1563	35151
HYPOAE AUDST AG-O D6-1014	CJ5 IIIB	12820	0.9297	29.0907	2753
HYPOAE AUDST CR-M D7-1116	MS0 IIIA	4288	1.4844	34.9427	3690	K5 VA, K0 VAB
HYPOAE SCRUA AA-A H145	WO0 I	<1	117.6641	8.5815	136				Unusual brownish colour
HYPOAE SCRUA AA-A H212	WN0 I	<1	101.3125	6.5448	11772	09 IIIA
HYPOAE SCRUA AA-A H212	O9 IIIA	<1	51.5156	185.5672	36653	WN0 I
IOCKOLT AA-A H0	O8 IIIA	<1	79.7422	145.2966	39111	WN0 I
iockolt aa-a h0	WN0 I	<1	79.7383	6.8117	49583	O8 IIIA
IOCKOLT AA-A H17	B9 IA0	<1	73.5742	483.6257	10273	B8 IA0, O0 VZ
iockolt aa-a h17	B8 IA0	<1	31.7852	449.1508	13207	B9 IA0, O0 VZ
MYRIERPH AA-A H153	09 IIIA	<1	99.0977	219.0652	33981
NYUENOOE AA-A H25	AEBE7 VAB	2	111.6133	0.9461	5406	BH	TTS1 VIx2, TTS8 VIx4, TTS6 VI, TTS2 VA, TTS4 VI, TTS5 VAB, TTS7 VAx2, TTS0 VI, TTS9 VI		A must see system!
NYUENOOE AA-A H120	B1 IIA	<1	75.9766	268.4974	53939.3	O0 VZ
OEPHAIRGH AA-A H14	O8 IIIA	<1	33.5547	180.0136	37454
ONGAIHN AA-A H1	B3 IA0	<1	103.7773	325.1876	25298	WC0 I, O0 VZx3
ongaihn aa-a h1	WC0 I	<1	43.6523	8.5162	2003	B3 IA0, O0 VZx3			Looks like a M class
ONGAIHN AA-A H24	B9 IA0	<1	76.1797	497.0230	10648	O0 VZx2
ONGAIHN AA-A H39	WO0 I	<1	84.6563	9.4320	61934
PHAA AOC AA-A H141	WNC0 I	<1	113.9180	5.2704	7105				Yellow-white - like F class
PHAA AUDST AA-A H7	B9 IA0	<1	47.8203	491.8203	10712	O7 IVA
phaa audst aa-a h7	O7 IVA	<1	38.0938	119.9152	42378	B9 IA0
PHAA AUDST AA-A H8	AEBE8 VI	2	39.4688	0.4856	5334	B0 VZx2, BH
PHAA AUDST AA-A H11	WO0 I	<1	66.6953	3.7523	5385				Yellow - like G class
PHAA AUDST AA-A H43	O9 IIIA	<1	59.4063	198.2641	35643
PHAA AUDST AA-A H75	AEBE3 VAB	2	109.5391	0.9430	5753	BHx2	TTS7 VAx2, TTS6 VA, TTS5 VA, TTS5 VB, TTS9 VI
PHAA AUDST AA-A H168	WO0 I	<1	107.6602	3.9982	42159	WN0 I
phaa audst aa-a h168	WN0 I	<1	60.0856	8.5551	56754	WO0 I
PHAA AUDST AA-A H176	WC0 I	<1	54.2227	3.4102	2416	AEBE3 VI, B0 VZx2, O0 VZ			Looks like M class
PHOI AOD AA-A H4	B9 IA0	<1	68.2461	497.5907	10221	O0 VZ, B0 VZ
PHOI AOD AA-A H31	WN0 I	<1	114.7500	7.2020	52613
PHOI AOD AA-A H181	WO0 I	<1	77.5117	9.9609	39430
PHOI AOD KT-Q E5-5036	K1 IAB	1284	2.3711	40.5116	4953	G3 VAB
PHUA AUSMS AA-A H23	WO0 I	<1	108.1953	6.8906	16073
PHUA AUSMS AA-A H48	WCO I	<1	59.5664	9.7250	60982	O0 VZ, AEBE VB
PHUA AUSMS AA-A H52	WNC0 I	<1	102.1797	4.5332	7890
PHUA AUSMS AA-A H85	O9 IIIA	<1	79.3320	188.1867	36544	O0 VZ
PLIMBEE AA-A H0	WO0 I	<1	113.9492	5.8774	11298	O8 IIIA
plimbee aa-a h0	O8 IIIA	<1	59.6016	167.7892	38030	WO0 I
PLIMBEE AA-A H20	WO0 I	<1	111.0078	3.5467	29547	WN0 I, O0 VZ
plimbee aa-a h20	WN0 I	<1	68.6328	5.7595	15087	WO0 I, O0 VZ
PUEKAAE AA-A H36	O9 IIIA	<1	89.0664	227.4564	33274	WC0 I, B3 IA0
puekaae aa-a h36	WC0 I	<1	55.0586	8.9514	43505	O9 IIIA, B3 IA0
puekaae aa-a h36	B3 IA0	<1	36.3359	331.1786	18767.7	O9 IIIA, WC0 I
PUELAA AA-A H11	WN0 I	<1	103.9844	9.8397	16617
STUEMAEA AA-A H95	WNC0 I	<1	115.0469	9.7096	44431	O8 IIIA
stuemaea aa-a h95	O8 IIIA	<1	70.4141	186.6754	36736	WNC0 I
THEEMEIA AA-A H83	WO0 I	<1	117.8047	3.8146	60070	WC0 I, O0 VZ, B3 IA0, WC0 I
theemeia aa-a h83	WC0 I	<1	45.6875	5.2422	24877	WO0 I, O0 VZ, B3 IA0, WC0 I
theemeia aa-a h83	B3 IA0	<1	36.4453	316.0911	25575	WO0 I, O0 VZ, WC0 I, WC0 I
theemeia aa-a h83	WC0 I	<1	43 5820	8.3745	20390	WO0 I, O0 VZ, WC0 I, B3 IA0
THEEMEIA AA-A H136	WNC0 I	<1	59.7461	4.5195	1615	B0 VZ			Red colour
THEEMEIA AA-A H188	WO0 I	<1	119.2031	6.4959	58237
WHAMBOO AA-A H30	O7 IVA	<1	80.5117	136.6751	40863	O7 IVA, WC0 I
whamboo aa-a h30	O7 IVA	<1	36.1719	122.3711	42080	O7 IVA, WC0 I
whamboo aa-a h30	WC0 I	<1	42.4414	4.8764	54327	O7 IVA, O7 IVA
WHAMBOO AA-A H32	AEBE3 VAB	2	56.2656	0.6270	4686	BH, O0 VZ	AEBE?, TTS?, TTS?, TTS?	TTS?, TTS?
WHAMBOO AA-A H33	O9 IIIA	<1	53.3203	219.1201	33744	O0 VZ
WHAMBOO AA-A H34	B8 IA0	<1	48.1250	455.028	13783

 

[Jackie Silver]

Great stuff, Pirin, that's a lot of extra data points for the big stars there. Thanks! Also one of the things I really want to find out is the evolution of the biggest stars that's all compressed into the "under 1 million years" bracket, and the progression from O to B forms a curve that some of these stars should fall on, so should be helpful for that in particular.

*grinning* A little while ago I tried to make a list of all the different topics of research that I'm trying to remember to do at the same time, and where we're at with each different area... ...it was a long list and I wasn't very succesful.

(edited to add) Pirin, if you get the chance would you look at the system map for PHOI AOD KT-Q E5-5036 and tell me if the star description says "Red Supergiant" or "Orange Giant" please?

 

[Esvandiary]

Just a quick note Jackie, before you end up very confused: it looks like your "sectors all" sheet might have incorrect widths; at least on the copy I imported into Excel, the numbers across the top go 62, 63, 64, 77, 78 ...

I don't have any access to the game for the next few days, so I'm trying to do more with what we've got. I've added some code to do individual class 1 runs (up to 35 in length), everything beyond that is quite simplistic and wrong for now.

 

[Jackie Silver]

Oops! Thanks, I will upload a corrected version shortly. I do make a lot of mistakes.

(edited to add)

Done. Corrected and updated version of the sectors all uploaded, also updated the log file with Pirin's data on big stars.

 

[Esvandiary]

Veering rapidly off-course: it appears (as I suspected) I still have a little work to do on the class 2 sectors. EDSM now has coords for one Graae Phroo KR-Y c3-1, which according to the current limited model is in a sector that shouldn't exist (at least not in the populated areas).

Confusingly, the sector next to it (Groa Eaeb) is absolutely correct, as are all the others in that row - if I change the data to make Graae Phroo match, I get a huge amount of "that's not right" from other data.
It should be a Phree, but it's a Phroo... At the moment, I have no idea why.

 

[Jackie Silver]

It should be a Phree, but it's a Phroo...

In the midst of the word he was trying to say, in the midst of his laughter and glee,
He had softly and suddenly vanished away, for the Phree was a Phroo-jum, you see!

 

[Esvandiary]

In the midst of the word he was trying to say, in the midst of his laughter and glee,
He had softly and suddenly vanished away, for the Phree was a Phroo-jum, you see!

Excellent.

Mystery possibly solved: I suspect the Phr suffix uses a restricted set of the suffixes, only actually going as far as "oo".

Interestingly, the list of these "short" prefixes in class 2 names is growing. More interestingly, the length of all of those "short" prefixes as they're used in class 2 is also within the range we have for the lengths of their class 1 runs...

Edit: I might be onto something... There are a few "oddball" start points in class 2 that didn't really follow the normal pattern, and I think they might not follow the normal pattern because the bits in the middle we can't see use short prefixes.