. /../StarMap2 discussion/ 12
night fth
written by Ferinex on Dec 18, 2008 03:00

:] Yuppo, starmap2 is done for all of you who use NoctisMapper.

cd/rars/Ferinex/starmap2 (317 Kb)
rose pony is best pony
written by Starchaser on Dec 18, 2008 04:04
Neuzd = Jesus of Felysia

Ferinix = Legend

Now I defiantly have to redownload Noctis and Noctis mapper.
r'lyeh sweet r'lyeh
written by Neuzd on Dec 18, 2008 19:34
Ferinex said:
:] Yuppo, starmap2 is done for all of you who use NoctisMapper.

cd/rars/Ferinex/starmap2 (317 Kb)
You mean this is a complete starmap2 built from the Noctis IV updated starmap?
A W E S O M E !
: D

And by the way, did you actually modified Triceratops source?
The Y axis suffers from the same old swapping problem, I never really realized where it originates, I guess in the infamous 'is_there' function
If you did, it would be nice to modify it further so that we can finally have a starmap2.bin with correct Y axis.

Just put a - in front of the Y parsis.
Check Balastrackonastreya to be sure:
-18928 ; 29680 ; -67336

-18928 ; -29680 ; -67336

[EDIT 2]
I found out "The Deepest Point" doesn't seem to be there. So I'm afraid that's not really complete...
2147069136 ; 726694 ; 2147149014
└> last changed by Neuzd on December 18, 2008 at 20:07
night fth
written by Ferinex on Dec 18, 2008 22:32
There's a problem with triceratops? I was never alerted of this fact.

So it flips the whole galaxy upside down, then? I could probably fix it, yes, but I need to know what it requires to compile properly. Dev-C++ and Visual Studio both fail.

EDIT: Bah, never mind. I'll simply edit the Source, then use the .bat to compile it. ;]

EDIT2: And it is created from the inbox that you released, updated, etc. Should be complete.
└> last changed by Ferinex on December 19, 2008 at 00:26
absent more than alex
written by Shadowlord on Dec 19, 2008 21:18
It was upside down (well, if y was the vertical axis)? Someone should have told me!

P.S. Could you kindly post the modified source (or upload it to cdisk) and let me know what you've changed so I can replace my copy here as well?
i do my own stun-- avatars
written by Albeyamakiir on Dec 20, 2008 08:04
The new map looks kinda cool. Wiggly lines curling round the centre and perfectly straight lines to the galactic edge.

By the way, I just noticed that the mapper clock is a few triads medius out. About 31 behind Noctis.
r'lyeh sweet r'lyeh
written by Neuzd on Dec 20, 2008 18:35
I just discovered another thing about Triceratops that bugs me a bit.

I don't really know if this is useful to report, but anyways....

It seems that Triceratops doesn't follow the entries in the StarMap.
I understand that parsing an entire galaxy of about 80 billions stars isn't a simple task, but I used the StarMap order listing to find out correct data about duplicate names.

It wouldn't be a great problem if the duplicate objects were of different types, but it happened that some of those objects were of the same type.
ARKHAM is the perfect example.

Both objects named ARKHAM are S00s and when I had my copy of Starmap2.bin generated by EvolveSM (which lists only stars) I was using their order in the starmap to retrive their Parsis coordinates.
The first being in StarMap was also the first being in StarMap2.

Now i noticed all records are messed up. FENIA should be the very first entry, but it isn't so in the starmap2.bin generated by Triceratops.

Ok, I don't really know if Shadowlord or Ferinex can (or want to) deal with this.
I'd be extremely grateful though, because this situation forces me to keep on using the EvolveSM generated starmap2.bin, and it's slow as hell....to find those 1300 stars I'm still missing will take a lot of time because it seems not to find the objects I'm looking for even when I set parsis to be really near the object I'm interested in.
lost, not forgotten
written by Alex on Dec 20, 2008 19:00
Neuzd said:
The Y axis suffers from the same old swapping problem, I never really realized where it originates
I'm not exactly remembering how is_there worked, but the original cause might be me, I suppose. I personally follow a convention whereas the Y axis grows as it goes DOWN, not up. Most people would call it upside down. For me, it's just because a framebuffer conventionally works by allocating scan lines and pixels from its top left corner, meaning that you'd have to change the sign of the Y coordinate of a dot before calculating an index to its pixel's location as L = W*Y + X.

In a 3D projection, however, that sign change could very easily be integrated in the multiplication by the distance from the focal plane, e.g.

Xs <- Xp * D / Z
Ys <- Yp * D / Z

could become:

Xs <- Xp * D / Z
Ys <- Yp * (-D) / Z

where (-D) could be a constant or variable. This would need no explicit instruction to change the sign of the Ys screen coordinate, later. So why am I still doing that? Well... it was useful in tracing 2D elements, many years ago, and I got used to having a reversed Y axis.

mmh... in next 3D engine I may fix this behavior of me (it's my own problem, not the engine's ). Still, I don't think I would easily tolerate adopting conventions where the Y and Z axes are swapped with respect to my habits, i.e. having a vertical Z axis and a "deep" Y axis.
absent more than alex
written by Shadowlord on Dec 21, 2008 08:25

There are a bunch of systems whose seed exists in multiple valid star system locations in the galaxy. I don't see ARKHAM in my list of them, but maybe it is a new one (Example: ? Is that what you meant, or did you mean someone or two someones tried to label two different systems of the same type with the same name?

About the starmap and starmap2 order differing: This is because, currently, Triceratops sends a star/body to the starmap2 file with fwrite as soon as it finds it while scanning the galaxy. (It loads starmap1 into memory, quicksorts it by seed, and then begins scanning and writing found objects to starmap2)

It's definitely possible to make it fix the order that it writes the starmap2 out in, it shouldn't be too difficult - say, don't write the objects to the starmap2 immediately, and read the starmap1 again after the whole galaxy has been scanned, and search the in-memory starmap2 database by seed for each starmap1 entry, writing out each starmap2 entry in order. Either Ferinex could do it, or I could (but he has made changes, yes? I don't have them. It would be nice if he would post them.).

Example duplicate system, if you're curious:
The code which outputs that is in the starmap2 -> starmap3 converter instead of Triceratops itself, so you would still end up with duplicate entries in starmap2 from Triceratops. This could be fixed fairly easily, though.

I posted about the duplicate systems way back in 2005 after I discovered it it:

About your missing systems:
At the top of triceratops.cpp, #define ANALYZE
and then recompile it and run it
it will be much slower but will tell you how many systems were omitted and why, so you can mark out some systems which are invalid or outside the galaxy (once upon a time there was no galactic distance limit besides what happened when you Went Too Far and Weird Things Happened). There are surprisingly many systems which are simply in invalid places, and it's rather odd that they're in the starmap, but perhaps people ended up in them due to bugs, or maybe some restrictions on system coordinates weren't originally in places in early NIV versions (just guessing on this one).

(I'd only use it with ANALYZE for determining whether it really is missing things)

Also, it won't notice labelled planets which are in unlabelled systems.
r'lyeh sweet r'lyeh
written by Neuzd on Dec 21, 2008 14:47
There are different cases of duplicates that can occur in Noctis, as we can see.

    - Total duplicates as those you're showing.
    - Anomalies like those stars being part of The Grid.
    - Totally DIFFERENT objects, with the same name.

My interest now is focused on the last type.
You see, those are legit systems, which represent a kind of problem while playing. It's a practical question about actually playing Noctis.

I'm gonna be brief and just illustrate an example. I'm also linking to documents I wrote about the subject.

There are 2 stars named Arkham in Feltyrion.
They are different stars in different locations AND they've been legitimely labeled in-game by 2 different people: Alessandro Roberti and Skinnymon.

    Arkham S00
    -46064 ; -40129 ; -1421011
    1 major body (ARKHAM I)

    Arkham S00
    191792 ; 6793743 ; -1527697
    9 major bodies, 18 minor (all named after Lovecraft's mythos)

This is a complete list of duplicate names (though a lot of parsis are missing), I'll compile a new one once I have a full starmap2.

This is an article I wrote about. It's comprehensive and includes everything I was too tired to repeat here (it's written in-character, just substitute Fenia with Alex and it shouldn't be difficult to understand).

Since on stramap2.bin there are no IDs, retriving the correct info for the duplicate is possible only if you know which of the 2 (or even 3, like DEATH) object you're interested in.
And this can be done only knowing its order in the starmap; if starmap2 has a different order (like Triceratops's) all my work goes in the WC: my goal is to finaly offer an online tool that can deal with duplicates and can show users those data that are impossible to be read again by ANY of the GOES modules.
I already have done a cool job, but the new starmap2 have puzzled all my plans of releasing it to the public this weekend.
doing pushups
written by Megagun on Dec 21, 2008 20:05
Megagun said:
Shadowlord said:
Megagun said:
*Skies of felysian and unstable (volcanic) worlds are bugged
*Volcanic worlds look wrong when viewed from space (this might also be the case for other planet types)
*Planet surfaces randomly change (volcanic worlds do that, and perhaps some others too)
*Random crashes happen, some without known causes, some with known causes (unstable planet + transparant surface = crash upon landing, for example)
When did those start happening?
Also, I forgot that WNICE has way more stars in the galaxy than DOSNICE..

Also, made some changes to Inbox32 so that it reports data and produces .log files.. It seems to import more labels than generic Inbox does... More investigation required (see also yesterday's Frespych logs)

Really, NICE currently has a few issues with it that are... well.. kind of sad..
night fth
written by Ferinex on Dec 21, 2008 20:05
Shadowlord said:
Either Ferinex could do it, or I could (but he has made changes, yes? I don't have them. It would be nice if he would post them.).
I haven't made any changes, no. I've been reading through the code and familiarizing myself with it first :] When I do start making changes, it's going to be more than just fixing the Y axis... I'm ging to make as many optimizations as I can, too, to try and speed it up.

So go ahead and make the changes, and send it my way. I'll work from the new code, so we don't have any forks.
lost, not forgotten
written by Alex on Dec 21, 2008 21:31
Oh, for reference, here's the converter that changes the sign of Y coordinates in starmap2.bin.

cd/zips/Alex/converter (24 Kb)

(and advertises L.in.oleum)
written by Ajax on Dec 21, 2008 21:34
Yes, we got it, Alex... thanks for the taunt on another Sunday...
lost, not forgotten
written by Alex on Dec 21, 2008 21:38
I've updated it a few minutes ago. Been adding a symbol hash and tagging letters to all variables.
reading this thread
no members are reading this thread
. /../StarMap2 discussion/ 12
48673, 13 queries, 0.115 s.this frame is part of the AnyNowhere network