RV Sonnenkreis - Decoding Universal Cartographics

[Macros Black]

What would be the easiest way to determine the number of systems in a sector; say PYRAEA EUR?

 

[Pirin]

...................

(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?

You hit the nail here Short answer: System map says Red Supergiant.

BUT....

There is a kind of mystery surrounding the procedure generated red K/M and blue A/B supergiants. On galaxy map all red supergiants appear as K I and all blue as B IA0. But all of them appear as M and A supergiants in the system map and when i target them with my ship when i am on supercruise inside the system. So basically i never know what star is that. Is it K or M? B or A? On other hand, it seems procedure generated M I, M I0, A I0 and Carbon C dont exist (although A I and B I do exist separately). Spent long time looking for those and never found a single one. Also dont recall any F II, F III, G II and G III. But never searched for those, so could be wrong.

In general it seems the script used to generate 99.999% of Galaxy used parameters for size, mass, temp, etc which are much lower that the ones which can be found in the hand made systems HD, HIP, etc. Some star classes where excluded altogether. You can never find anything close to Eta Carinae in the procedure generated space. The size of all giants and supergiants seem to be in 28 - 5X range. The only stars which FD allowed to grow bigger than that are O III and O IV which seem to have limit of about 230 and B (A?) supergiants which have a limit of 500. The mass limit for all stars seem to be 120.
Something obviously went wrong when Star Forge was creating WRs. Some of them have abnormally low temperatures. Hypoae Scrua AA-A H145 is especially amusing in that aspect.

It also seems that in HD and HIP systems not only the main stars were hand made. Some of the planets have values for temp, mass and pressure far above anything that can be found elsewhere.

 

[Jackie Silver]

What would be the easiest way to determine the number of systems in a sector; say PYRAEA EUR?

Good question, it would depend heavily on where the sector is. I think you would need to look at the broad shape of the sector, then split it up into representative homogeneous chunks, take samples within each chunk of different mass codes and multiply by the number of those mass codes within a sector. I can have a go at it if you like, it would certainly be an interesting exercise?

There is a kind of mystery surrounding the procedure generated red K/M and blue A/B supergiants. On galaxy map all red supergiants appear as K I and all blue as B IA0. But all of them appear as M and A supergiants in the system map and when i target them with my ship when i am on supercruise inside the system. So basically i never know what star is that. Is it K or M? B or A? On other hand, it seems procedure generated M I, M I0, A I0 and Carbon C dont exist (although A I and B I do exist separately). Spent long time looking for those and never found a single one. Also dont recall any F II, F III, G II and G III. But never searched for those, so could be wrong.

Absolutely, it's strange. We could look at the temperature data to see whether the galaxy map spectral class or the system map description better matches the star's actual temperature; with A < 10000K and B > 10000K it should be doable for the big supergiants at least. So far as I can tell (there's some diagrams upthread where it stands out) there are broadly two distinct ranges of giants - the more luminous group are all very young (<1 MYr) and huge and the less luminous group are older, smaller evolved giants. But there are subgroups within the less luminous group, and the more luminous group are joined by O and B stars evolving just off their very short main sequence.

Something obviously went wrong when Star Forge was creating WRs. Some of them have abnormally low temperatures.
Some of the planets have values for temp, mass and pressure far above anything that can be found elsewhere.

Yes and yes. There are also some (a very few) stars in the Bubble which have galaxy map spectral classes suggesting they are proc-gen, but data suggesting they're catalogue stars. IIRR Lave or Leesti is one.

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.

Cool - I still haven't got my head round those class 2 ones.

 

[Esvandiary]

Cool - I still haven't got my head round those class 2 ones.

I'm not surprised - there's still a whole lot of "manual" data entry (we're basically specifying what the start system name should be for each prefix run, just in quite a compact form).
Perhaps I shouldn't have stopped working on them when I did - although I suppose I probably wouldn't have got much further without your observation about prefixes having a specific length.

Anyway, there is light at the end of the tunnel - I now have a generator that can start from the very bottom-left of the galaxy and produce class 2 sector start points ad infinitum (assuming the correct prefix run lengths, of course). Only managed that after deciphering two additional layers of checkerboarding, naturally... [wacky]

Edit: Forgot to say, first prediction: either Oo or Eu are 31 in length (I guess more accurately, they add up to a combined 66). For some reason my money's on Eu being the short one, but I can't remember why I thought that.

Edit 2: More predictions, I reckon Kyl is 30 long and Hyp is 25...

 

[Macros Black]

Good question, it would depend heavily on where the sector is. I think you would need to look at the broad shape of the sector, then split it up into representative homogeneous chunks, take samples within each chunk of different mass codes and multiply by the number of those mass codes within a sector. I can have a go at it if you like, it would certainly be an interesting exercise?

Cool, I like

 

[Jackie Silver]

I'm not surprised - there's still a whole lot of "manual" data entry (we're basically specifying what the start system name should be for each prefix run, just in quite a compact form).
Perhaps I shouldn't have stopped working on them when I did - although I suppose I probably wouldn't have got much further without your observation about prefixes having a specific length.

Anyway, there is light at the end of the tunnel - I now have a generator that can start from the very bottom-left of the galaxy and produce class 2 sector start points ad infinitum (assuming the correct prefix run lengths, of course). Only managed that after deciphering two additional layers of checkerboarding, naturally... [wacky]

Edit: Forgot to say, first prediction: either Oo or Eu are 31 in length (I guess more accurately, they add up to a combined 66). For some reason my money's on Eu being the short one, but I can't remember why I thought that.

Edit 2: More predictions, I reckon Kyl is 30 long and Hyp is 25...

The more lengths I get close to, the more it seems that the longer a phoneme is, the more likely it is to be short. Not certain, but it looks like all the single letter phoneme prefixes are sequences of length 35, while the likes of Pyth, Aea, and Lych are short.

Macros, I'll have a look at that sector and see what I come up with - may take a while. It may be more difficult as it's on the outer rim of the galaxy.
At first look, I've noticed something very interesting about the distribution of stars - the mass code B systems are largely (entirely?) confined to the plane (about 60ly height, or 3 B mass codes high), C systems extend higher (about 440ly height, or 11 C mass codes high), D seem to be unconfined (but very sparse), highest mass code is E. This is fascinating especially if the same pattern exists closer to the Core (but is obscured by the sheer number of stars and the limits being bigger than a sector.) Could explain how the Forge throttles star placement!
I'm busy jonking towards TZAIWNS at the moment but will come back to this later.
There's an interesting little chain of stars near PYRAEA EUR TO-Q D5-0.

 

[Esvandiary]

The more lengths I get close to, the more it seems that the longer a phoneme is, the more likely it is to be short. Not certain, but it looks like all the single letter phoneme prefixes are sequences of length 35, while the likes of Pyth, Aea, and Lych are short.

I noticed that a bit last night, wasn't sure if there was anything to it!

Well, my generator is now nearly getting as far as Z=0 before finding something that the existing methods disagree with... Onwards!

Edit: And now it can do the whole way! \o/

 

[Jackie Silver]

Here's an estimate for the number of star systems in PYRAEA EUR.

I've made the following simplifying assumptions based on a look at the way stars are distributed in the sector:

Mass Code A systems are not present in significant numbers (if at all) and can be discounted.
Mass Code F,G,H systems are not present.
Mass Code E systems are present but in such small numbers that they can be discounted.

Mass Code B systems are present across the lowest level of the sector, in a cuboid of 64 x 64 x 3 subsectors.
Mass Code C systems are present across the lower half of the sector, in a cuboid of 32 x 32 x 11 subsectors.
Mass Code D systems are present across the whole sector in a cube of 16 x 16 x 16 subsectors.

I picked a number of reference subsectors spread evenly across the sector and half-way through the relevant cuboid or cube. I counted the number of systems within each subsector:

VT-Z B13 - 1 system
AV-Z B13 - 1 system
XO-X B28 - 1 system
PC-X B42 - 1 system
UD-X B42 - 1 system
Average number of systems in a B subsector - 1

ZP-O C6 - 1 system
OQ-O C6 - 1 system
NN-A C14 - 1 system
HE-O C20 - 2 systems
WE-O C20 - 1 system
Average number of systems in a C subsector - 6/5

BB-W D2 - 2 systems
FB-W D2 - 2 systems
VZ-O D6 - 1 system
DS-J D9 - 3 systems
KS-J D9 - 2 systems
Average number of systems in a D subsector - 2

This gives a total of:
B - 3 * 64 * 64 * 1 = 12288
C - 11 * 32 * 32 * 6/5 = 13517
D - 16 * 16 * 16 * 2 = 8192

Around 32,000 systems in total.
At a ballpark 2 stars per system, call it 64,000 stars*.

It will be interesting to see how the structure of this sector compares to others - we might be able to get some good insights into overall galactic structure.

*this is worryingly within the reach of a suitably fanatical explorer, Macros.

 

[Esvandiary]

Right, the fully-algorithm-ised class 2 code is pushed. I've left the old semi-manual-data code in there for now, just to allow for easy testing as we tweak the numbers for class 1.
For reference, from a naming point of view the galaxy appears to be 128 x 16 x 78. With some adjustments for the not-full-length sectors, the class 2 names slotted nicely into place.

Speaking of which, I'm really struggling with the continuations on class 1... I've got the individual runs down fine, but I have no idea how the wider continuations happen. My concern is that due to the nature of the class 1 sequences (lots of prefixes in a row and mostly long runs), we might not have that many examples to work from. I'm not 100% convinced I've actually seen any so far...

Edit: Looks like something is still a bit wrong... Taking a look at it.

 

[Jackie Silver]

Edit: Looks like something is still a bit wrong... Taking a look at it.

That's my line! I will continue to look at getting more exact sector lengths - I'll concentrate on the middle band of the -1 plane where the opportunities are good.

Also jonking onwards, around 9 Kylies to go to TZAIWNS. I have no idea what I'll look for when I get there and it's almost certainly a wild goose chase. There's a certain irony in that not so long ago I found the trip to Heart & Soul a long one, and yet now I give myself a 20 Kylie detour without a second thought.

 

[Esvandiary]

That's my line! I will continue to look at getting more exact sector lengths - I'll concentrate on the middle band of the -1 plane where the opportunities are good.

Also jonking onwards, around 9 Kylies to go to TZAIWNS. I have no idea what I'll look for when I get there and it's almost certainly a wild goose chase. There's a certain irony in that not so long ago I found the trip to Heart & Soul a long one, and yet now I give myself a 20 Kylie detour without a second thought.

More exact sector lengths would be very helpful indeed. I think I've nailed down quite a few as part of the C2 set (although not all of those in that list I'm sure about - many of them have come from your spreadsheet estimates... I'll separate them out soon); I've also now fixed the depressingly small error that was causing the EDDB test to quite rightly go haywire, so all should now finally be happy in class 2 land.

Haven't got any further on the remaining problem, though...

Best of luck finding whatever you're hunting in Tzaiwns!

 

[Jackie Silver]

More exact sector lengths would be very helpful indeed.

I got another couple, Ou is 35 and Sk is 27. I've uploaded an updated version of the list. I'm getting quite a lot of them narrowed down closer to 35.

 

[Jackie Silver]

Google Docs version of the sector map.

Drop me a PM anyone who'd like to edit. I'd like to keep it fairly tightly controlled, mind, because honestly I know how often I've screwed it up while entering data and mistakes have the potential to snowball while still going unnoticed for a long time, which would be bad. So please please please be careful to get the plane, position and spelling right! [heart]

(Also I gather the planetary mining database was vandalised. Not cool.)

Another length: Hy is 35.

Motoring. Less than a Kylie from my destination.

 

[Esvandiary]

Woah, where's our favourite meerkat gone?!

Also, this is proving... challenging.
Thought process:

"Hm, OK, so it looks like it goes through the suffixes, and increments phoneme 3 if it runs out. Then when it finishes that prefix run, it increments the prefix and starts the other phonemes where the previous run started."

... --> Choomaae --> Choomeau --> Choomee --> Myriesly --> Myrielk --> ...

"Right, except the non-full-length ones seem to have their own pattern; that makes some sense, would be tricky to somehow keep them synced with the full length ones. OK, I think I'm starting to get a handle on this, maybe soon I-"

... --> Myrierph --> Cleeque --> Cleeqai --> ??? --> Cloomoi --> ...

"... Well 'q' isn't that close to 'm' in the infixes list... OK, so maybe 'q' as an infix has its own restricted run of suffixes that ends at 'ai', after which it resets to anoth-"

EDI> find -r '^\w+q\w+ ..-. .+$'

Matching systems:

  Buqoa BV-X d1-0
  Cleeque HX-L d7-1325
  Cleeque JM-W d1-122
  Cuqe ZK-P d5-944
  Cuqe ZK-P d5-996
  Dryiqau VL-J d10-33
  Dryiqeau GS-R d5-20
  Dryiqi IF-R d4-11
  Dryiqia CU-W a97-1
  Dryiqia HG-J b1-1
  (50+ more)

"..." *tableflip*

 

[Jackie Silver]

Woah, where's our favourite meerkat gone?!

I left him at a place called Lee Ho Fooks...

I've mostly been occupied traveling so haven't made much progress apart from filling in more of the sector map. I reached TZAIWNS, realised the enormity of the task involved in searching the sector and settled for a leisurely stroll between some of the brighter stars, paying more attention as I went. This paid off with finding several ELW, but nothing else unusual. I'd been wondering if the sector code (being single length as TZ is) might have some effect on the stars in the sector, but it doesn't seem to. Possibly it would have an effect if the sector wasn't so far out; there aren't any big or unusual stars in TZAIWNS. I've looked for another TZ nearer to Sadge, but I may visit a PYTH one instead.

(In orbit over a water world; will you look at that grime!)

Incidental music: Lorde, The Long Blondes

Edited to add: G, H, I, Sp are all almost certainly 35. Should have more soon, not quite at the point where everything snowballs to completion but getting there.
Also: A, Sh, B, C, D, all 35. M, St, N also 35.

 

[Esvandiary]

OK, current theory...

It iterates over the prefixes strictly in order, wrapping as needed (let me know if you can spot any cases where this doesn't appear to be correct).
Within each run of prefixes, it iterates over the suffixes from a given start point (not necessarily the start of the suffix list). If it wraps back over its starting point, it increments phoneme 3 and continues (no idea yet what happens with three-phoneme names).
e.g. ... --> Veqia --> Veqie --> ... (many) ... --> Vegnoi --> Vegnaa --> ... (many) ... --> Vegnao --> Vegnau --> Weqo --> Weqeae --> Wequ --> Weqaea --> Weqia --> Weqie ...
Starting point in this instance seems to be "o", and the distance between Veqia and Weqia is 35 as we might expect.

At some point (seemingly not when it wraps back to the start of the prefix list, but not sure as I'll describe later) it increments phoneme 2. This doesn't appear to "reset" the rest of the sequence.

From what I can see from looking at sector names as they go "forwards" in Z at the same height (i.e. having completed one full Z slice), each slice appears to get through ~35-55 prefixes; naturally it changes a bit depending on which prefixes it goes through since they can have different lengths.
The big problem, however, is that as far as I know, we have no way of knowing for sure that it's getting through X prefixes, rather than (full list + X) prefixes. Since how/when it "resets" is still something of a mystery, that's still somewhat up in the air.

I made a quick mode of my script to test how many prefixes it goes through in a Z slice (and in theory how many total sectors, if it's doing a full run through each one on the way).
The results, at X=0, Y=0 (i.e. going through Wregoe and Stuemeae):

Spoiler

λ python pgnames.py pdiff Spl V Gr Sp Sch My Ea Gl H Ou Wr S Phyl Sm Tr Syr Lys Pr Sl Eo Syn P Fr Io Oe Wh O Shr D Z Myr St Ph B Ee Ly L Ai Spl V Gr J Hyp Sty Ea Gl H Ou Wr S Tyr Sm Tr Syr Q Pr Sl Oo Syn P Fr Io
Spl --> V: 38 prefixes (rollover: False, predicted len: 1187)
V --> Gr: 47 prefixes (rollover: True, predicted len: 1205)
Gr --> Sp: 42 prefixes (rollover: False, predicted len: 1176)
Sp --> Sch: 42 prefixes (rollover: True, predicted len: 1207)
Sch --> My: 49 prefixes (rollover: False, predicted len: 1198)
My --> Ea: 40 prefixes (rollover: False, predicted len: 1212)
Ea --> Gl: 47 prefixes (rollover: True, predicted len: 1205)
Gl --> H: 42 prefixes (rollover: False, predicted len: 1176)
H --> Ou: 40 prefixes (rollover: True, predicted len: 1216)
Ou --> Wr: 51 prefixes (rollover: False, predicted len: 1189)
Wr --> S: 40 prefixes (rollover: False, predicted len: 1212)
S --> Phyl: 45 prefixes (rollover: True, predicted len: 1202)
Phyl --> Sm: 44 prefixes (rollover: False, predicted len: 1179)
Sm --> Tr: 39 prefixes (rollover: True, predicted len: 1212)
Tr --> Syr: 52 prefixes (rollover: False, predicted len: 1197)
Syr --> Lys: 39 prefixes (rollover: False, predicted len: 1198)
Lys --> Pr: 44 prefixes (rollover: True, predicted len: 1186)
Pr --> Sl: 46 prefixes (rollover: False, predicted len: 1205)
Sl --> Eo: 38 prefixes (rollover: True, predicted len: 1181)
Eo --> Syn: 53 prefixes (rollover: False, predicted len: 1228)
Syn --> P: 38 prefixes (rollover: False, predicted len: 1171)
P --> Fr: 44 prefixes (rollover: True, predicted len: 1210)
Fr --> Io: 46 prefixes (rollover: False, predicted len: 1173)
Io --> Oe: 39 prefixes (rollover: False, predicted len: 1216)
Oe --> Wh: 52 prefixes (rollover: True, predicted len: 1195)
Wh --> O: 38 prefixes (rollover: False, predicted len: 1193)
O --> Shr: 44 prefixes (rollover: True, predicted len: 1225)
Shr --> D: 47 prefixes (rollover: False, predicted len: 1177)
D --> Z: 39 prefixes (rollover: False, predicted len: 1208)
Z --> Myr: 51 prefixes (rollover: True, predicted len: 1209)
Myr --> St: 39 prefixes (rollover: False, predicted len: 1179)
St --> Ph: 42 prefixes (rollover: True, predicted len: 1196)
Ph --> B: 49 prefixes (rollover: False, predicted len: 1206)
B --> Ee: 39 prefixes (rollover: False, predicted len: 1208)
Ee --> Ly: 48 prefixes (rollover: True, predicted len: 1188)
Ly --> L: 41 prefixes (rollover: False, predicted len: 1199)
L --> Ai: 41 prefixes (rollover: True, predicted len: 1188)
Ai --> Spl: 50 prefixes (rollover: False, predicted len: 1198)
Spl --> V: 38 prefixes (rollover: False, predicted len: 1187)
V --> Gr: 47 prefixes (rollover: True, predicted len: 1205)
Gr --> J: 43 prefixes (rollover: False, predicted len: 1211)
J --> Hyp: 42 prefixes (rollover: True, predicted len: 1207)
Hyp --> Sty: 49 prefixes (rollover: False, predicted len: 1198)
Sty --> Ea: 39 prefixes (rollover: False, predicted len: 1177)
Ea --> Gl: 47 prefixes (rollover: True, predicted len: 1205)
Gl --> H: 42 prefixes (rollover: False, predicted len: 1176)
H --> Ou: 40 prefixes (rollover: True, predicted len: 1216)
Ou --> Wr: 51 prefixes (rollover: False, predicted len: 1189)
Wr --> S: 40 prefixes (rollover: False, predicted len: 1212)
S --> Tyr: 46 prefixes (rollover: True, predicted len: 1210)
Tyr --> Sm: 43 prefixes (rollover: False, predicted len: 1171)
Sm --> Tr: 39 prefixes (rollover: True, predicted len: 1212)
Tr --> Syr: 52 prefixes (rollover: False, predicted len: 1197)
Syr --> Q: 40 prefixes (rollover: False, predicted len: 1208)
Q --> Pr: 43 prefixes (rollover: True, predicted len: 1176)
Pr --> Sl: 46 prefixes (rollover: False, predicted len: 1205)
Sl --> Oo: 39 prefixes (rollover: True, predicted len: 1216)
Oo --> Syn: 52 prefixes (rollover: False, predicted len: 1193)
Syn --> P: 38 prefixes (rollover: False, predicted len: 1171)
P --> Fr: 44 prefixes (rollover: True, predicted len: 1210)
Fr --> Io: 46 prefixes (rollover: False, predicted len: 1173)

The good news first: the number of prefixes it goes through - and the number of sectors that would entail - is relatively consistent. A bit of variation as we'd expect from the different run lengths used by different prefixes.
The bad news: those sector length estimates are nowhere near the number of sectors in a Z slice (128 wide x 16 high = 2048 sectors per slice). Hum.

 

[Jackie Silver]

It iterates over the prefixes strictly in order, wrapping as needed (let me know if you can spot any cases where this doesn't appear to be correct).

It's been correct everywhere that I've seen, to the extent that if I get one out of sequence I know that I've made a mistake.

I'm trying to follow the wrapping between levels (from 'down' to 'up' along the sheet) and that seems to be working exactly using the 128 size grid - I can count along from a known position on one of the lower levels and follow it through hundreds of sectors to another known position on the upper levels. Could it be repeating over 10 up / down slices? (1280 sectors from "down-left(west)" to "up-right(east)") That would more-or-less fit the rotation as far as I can see?

There was an interesting throwaway comment somewhere recently along the lines of "well, we've got a cluster way above the galaxy that we like the view from"... ^^

(eta)

I got a bunch more (same link) but I'm less certain that SPL is 17 - it could be 16 after all - it's either that or one of the others between HY and SPL is shorter than expected.

 

[Ziggy Stardust]

Trying to get to grips what you fellers are on about. Trying and failing spectacularly.

But it all sounds quite marvelous! Keep up the good work chaps!

And yeah, I do miss the meerkat.

 

[Esvandiary]

Trying to get to grips what you fellers are on about. Trying and failing spectacularly.

But it all sounds quite marvelous! Keep up the good work chaps!

And yeah, I do miss the meerkat.

Yeah, I think this is one of those things that you're either very engrossed in, or you're sat there going "what the ... are they talking about"

The meerkat returns!
I don't think we had anything against what you changed it to, Jackie - just that our furry friend is very iconic, a good fit I think. Simples!

Either way - those predicted lengths earlier at least tell us we're on the right track with the run lengths; they might not be around the magnitude we'd have liked, but they're really quite consistent.
I'll give it a try with a height of 10 and see what I can come up with... It feels very odd to me that they'd use a different height for one naming scheme to the other though; I'm wondering if there's something else going on that we're not accounting for yet. Or they just found the one-word pattern ran out too soon, got lazy, and restricted the height.

Jackie, when it's next convenient, any chance I could get an updated prefix list spreadsheet with the min/max run lengths? Would be good to see the latest for where you're up to, and correct my values where needed.

 

[Jackie Silver]

Here - sorry, I meant to post that earlier. I'll put that on to Google Docs as well sooner or later. I made a copy of the big map and started filling in the sequences on it, was making a lot of progress getting the lengths but it's very messy so I'll start over again with it sometime. I don't want to mess up the online version with all that stuff especially when it's speculative.

Trying to get to grips what you fellers are on about. Trying and failing spectacularly.

Imagine the galaxy is a giant slice of Battenberg which for reasons beyond our ken has had small chunks of carrot cake pushed into it all over the place...

Currently jonking on a course homewards that will skirt the outer edges of the Core. Moving fast for a short-legged Python.