New Celestia unified runtime database

[Lafuente_Astronomy]

A quick glance through tokenizer.cpp indicates the only time '[' or ']' is parsed is during value assignment for arrays.

Multinaming is also cross indexing, this is part of the catalog issues being discussed.


Janus.

Well, I'll just be patient until there are fixes to the coding itself. I can give suggestions with regards to catalogs if you want, since I posted a Wikipedia page of the list of Astronomical Catalogs but I can do it again here: https://en.wikipedia.org/wiki/List_of_astronomical_catalogues

[gironde]

If you want to modify the reading mode and its interpretation by the Celestia engine, you must make the name of the reference catalog, ie HIP and TYC, mandatory.
rather than write

Code: Select all

32349: Sirius: Sirius A: GJ 244 A: 2MASS J06450887-1642566 A: 9 CMa: BD-16 1591: HD 48915: HR 2491: IRAS 06429-1639: LHS 219: SAO 151881: SD-16 1591: WDS J06451- 1643 A: ALF CMa: ADS 5423

we will have to write

Code: Select all

HIP 32349: Sirius: Sirius A: GJ 244 A: 2MASS J06450887-1642566 A: 9 CMa: BD-16 1591: HD 48915: HR 2491: IRAS 06429-1639: LHS 219: SAO 151881: SD-16 1591: WDS J06451 -1643 To: ALF CMa: ADS 5423

and in the stc file:

Code: Select all

Star HIP 32349 # Sirius or Sirius A starnames.dat
{
OrbitBarycenter "Sirius A-B"

SpectralType "A1V"
AppMag -1.43

EllipticalOrbit
{# fully specified orientation
period 50.09
SemiMajorAxis 6.73 # mass ratio 1.99: 1.03
eccentricity 0.5923
inclination 97.51
AscendingNode 161.33
ArgOfPericenter 4.56
MeanAnomaly 40.89
}
}

--------------------------------------------------------------------------

ps: the use of "/" in the name of a barycenter (or a star) is to be avoided because Celestia interprets this character in a particular way. it is better to use "-"

Code: Select all

Barycenter "Sirius A-B"

[Lafuente_Astronomy]

That looks good. Perhaps it can be one of the ways to solve the starnames.dat limitations. But should the same be included for any other catalogs and their identifiers, i.e GAIA, 2MASS, HD, SAO, etc, or should it be just limited to HIP and Tycho instead?

[Janus]

One of the things I would reccomend is that star:name & star:designation both be arrays/vectors/lists of strings.

Let a star have as many names and designations as it wants.
They can all be preindexed, and still be available for any presentation style.
Have a keypress, dropdown, popup, or script command to change the index of the displayed name or designation in mass, or by star.
If you are doing a presentation, have a keypress that cycles the displayed names.
Focus on a single star, then show the names and asterisms of various cultures, without changing the view.

Just some thoughts.


Janus.

[Lafuente_Astronomy]

One of the things I would reccomend is that star:name & star:designation both be arrays/vectors/lists of strings.

Let a star have as many names and designations as it wants.
They can all be preindexed, and still be available for any presentation style.
Have a keypress, dropdown, popup, or script command to change the index of the displayed name or designation in mass, or by star.
If you are doing a presentation, have a keypress that cycles the displayed names.
Focus on a single star, then show the names and asterisms of various cultures, without changing the view.

Just some thoughts.


Janus.

Looks like a good idea. More designations should be added as per their memeberships in those different catalogs

Added after 8 minutes 50 seconds:
Perhaps the naming arrangement can also be done like Stellarium, in which its traditional names are presented first, then below those names are a series of identifiers. For example:

Betelgeuse (Betelgeuze - Betelgeux - Al Mankib) (Add possible traditional and cultural names for Betelgeuse)
α Ori, 58 Ori, HR 2061, BD+7°1055, HD 39801, FK5 224, HIP 27989, SAO 113271, GC 7451, CCDM J05552+0724AP, AAVSO 0549+07 (Add other catalog identifiers for Betelgeuse)

That way, the names and identifiers can be properly ordered in Celestia

[Janus]

Designations as numbers can also be handled the same way.
We start with an enum listing all the acecpted catalogs.
Then an array of strings indexed by the enums.
When parsed, the designator is the text upto the first space, the number taken from the rest.
While a uint32_t would be fast in one way, it only works for numeric only designations, and there is more hashy way.

As long as you are only going for an exact match only, store the md5 hash of the text name of an object.
Generate the hash as they type, and search for it.
Does not apply to partial matches though.
Unless you want to generate a hash every four or so characters, which is memory intensive, and overkill.

This is all just throwing ideas out to see what people think of view points.


Janus.

[Lafuente_Astronomy]

Perhaps that might do as well. But while they can be searched as per their identities, obviously it would be better to search those stars by their familiar names, because it's easier and faster that way. The catalog names will just be there for identification's sake

[Janus]

After thinking about it for a while, I changed my mind somewhat.
We start with an enum of catalogs.

Celestia = 0,
Hip = 1,
Tyc = 2,
????

We then have an object/class with two members.

string ObjectName;
[array of uint64_t] Indexes;

The first is a name, any name.
The second is an array 64 bit integers which are the numeric indexes in catalogs.
The first index is filled in at load time, an in memory index or pointer as preferred.
The rest of them are catalog numbers, with zero meaning "not in this catalog".
In this way, all catalogs could have a single interface, without caring if they were stars, DSOs, planets, or whatever.
Add a celx command for childof() and get the ability to select/enumerate based on parentage.
Returning a table to celx is simple, and this way the function does not care what it returns.

You could have catalogs of asteroids, comets, black holes, rogue objects, whatever you wanted.
All with a singular interface, with things like shorthand expansion as an option.

Or you could add a third member, ObjectData which is a pointer filled in at load time, per catalog.
I realize that is somewhat redundant with Indexes[0] pointing to the loaded structure, but there may be uses I have not thought of.


Janus.

[Lafuente_Astronomy]

You could have catalogs of asteroids, comets, black holes, rogue objects, whatever you wanted.
All with a singular interface, with things like shorthand expansion as an option.

That's the original idea with the proposed Celestia Catalog as the Universal Identifier Database. The Celestia Catalog as a whole contains ALL objects regardless of type and characteristic, but are subdivided by type. The purpose of the Singular Celestia Catalog would then be both for identification and search purposes

Added after 1 hour 16 minutes:
Well, here's an idea that can be done: While the Celestia Catalog would be all-encompassing all objects so that they can be well-identified in search, there must be subdivisions of that catalog by making smaller catalogs that could make the searches more specific without breaking and ruining the overall search capabilities of the entire Celestia Catalog.

The Celestia Catalog would be subdivided into several smaller catalogs for more specific searching and identification. And it is through typing the specific catalogs that we can narrow it down to any specific object that you're searching for.

So for example:
One of the subdivisions of The Celestia Catalog should be the Celestia Catalog: Galaxies. Identification 0 should be none other than the Milky Way Galaxy because it is our home galaxy.
Under that specific subcatalog of The Celestia Catalog are smaller, more specific subdivisions of The Celestia Catalog. Like for example, Celestia Catalog: Identification 0(Milky Way) Stars. In that catalog, Identification 0 would then refer to none other than our own Sun.
Coinciding with The Celestia Catalog: Identification 0 Stars is The Celestia Catalog: Identification 0 SSO, meaning Solar System Objects, covering all non-stellar objects orbiting star systems, from planets to asteroids, to members of debris disks, etc. But each member would have a number based on its parent star and a different prefix based on their characteristics as an SSO.

So, let's search for Earth in that Catalog. Here's how it would appear in the catalog:
CELG 0/CELG 0S 0/CELSSO 0P3
CELG refers to "The Celestia Catalog: Galaxies" with the Milky Way as Identification 0, CELG 0S refers to "The Celestia Catalog: Identification 0 Stars", referring to all stars in the Milky Way(Identification 0), and the Sun as that catalog's Identification 0. Then CELG 0SSO 0P3 refers to "The Celestia Catalog: Identification 0 Solar System Objects with "0P3" referring to the third planet of Identification 0(The Sun), which is none other than the Earth itself.

Another example, this time, we search Proxima Centauri b
The result would go like this:
CELG 0/CELG 0S 1/CELSSO 1P1
I identified Proxima Centauri b as "1P1", because the catalog should be numbered and identified via distance to the Sun itsef, so the first hundred or thousand stars will be of stars that are around 1-100 LY away from the Sun. Since Proxima Centuari is the closest star to the Sun, then it takes the Identifier 1, with Rigil Kentaurus A and B taking up 2 and 3 respectively. And the P1 is because assuming Proxima Centauri b is the first planet of the Proxima Centauri system, then it is then, the first planet as stated in the catalog.

However, if you devs and programmers have a better idea, let me know. These things are after all, just my 2 cents

Added after 4 minutes 14 seconds:
For asteroids, they should be numbered and identified by the distance of their perihelion to the Sun. So, since there exists an asteroid called 1566 Icarus: https://en.wikipedia.org/wiki/1566_Icarus, then Icarus takes up the identity of the first asteroid in the catalog.
As shown here:
CELG 0/CELG 0S 0/CELSSO 0A1
Obviously the "A" means asteroid. And they would be cataloged and identified regardless of their class, be it Near-Earth Asteroid, Asteroid Belt Asteroids, Jupiter Trojans, etc.

[Janus]

The very first thing I apply to any program/troubleshoot/build/repair/diagnose issue/problem is occam's razor.
The simplest answer is usually the right one.
I bring this up because of the plethora of ideas presented here.

The simplest way I can put how I see it is like this.
What is the simplest combination that fits the data?

One that I explain to customer after customer is actually kind of silly.
Binary data is for computers, text is for people.
Only make a binary file complicated if you really have to.
The same applies to text, and especially to text.
Everything shown has to be presented to be read, and skimmed over or scanned.

If the end user/reader can not make sense of the structure at a glance, it is to complicated, whether the data is messy or not.
Sometimes this is necessary, or at least unavoidable, which stinks.
Sometimes data just won't allow itself to be displayed neatly, though mostly that is a programmer problem.
The presentation of a structure should guide the eyes, not make a jigsaw or picture puzzle.

Celestia displays other sources of data, it is not a source, just a display engine, it must not self reference.
An astronomer should be able to type in a designation in parent child order, starting where ever.

Milkyway : Sol : 86 Seleme for instance.
Sol : 112 Iphignia would also work.
Galaxy 0 : Star 0 : C2012 S1 - Ison is another.

This is also the way text files should be formatted as well.
You need clear demarkers that do not have other jobs like either slash / \ does.
I use a colon, normally with a space around it, because it is a common demarker not seen in names like even the comma is sometimes.
The spaces around it help the eye see the demarkations, and speed searches because you can see how many " : " occur to count layers.
The colon is also a common demarker in cvs data, which spreadsheets all accept.
I like the idea of creating a script driven data file for later analysis.

I prefer separate databases in memory to aid drawing the universe in layers.
DSOs, Galaxies, Stars, Planets, you get the idea.
I do not believe that even with indexing, the one size fits all solution will scale.
I also freely admit I could be wrong.


Janus.

[Lafuente_Astronomy]

Alright. That could work. But let's see how the devs will take on this and either apply it to the program or make an entirely new one altogether

While my idea is not well-written in code, the idea is still there. One has to make a specific catalog that contains all catalogs within it, and these catalogs contain catalogs within it in turn,so as to give a proper hierarchy and set for objects in Celestia. All Milky Way Objects belong to the Milky Way's catalog, all Solar System Objects belong to the Solar System's catalog, and so on. The gist of the idea I present is more or less, dividing up the absolute Universal Object Identifier(The Celestia Catalog) into several divisions, with each division being divided into several smaller divisions, all for specification's sake. I, for one, thinks that it's a very interesting take, as all objects in Astronomy can be categorized and be placed in hierarchal divisions, from the largest to the smallest

[Janus]

Using any hierarchy other than the one nature/reality uses seems silly.

If you want a simpler way to look at what I see, try this.
It is like the difference between the foot thick dictionary, and a set of encyclopedias.

While the one is nice from a single source point of view, the collection is more precise and adaptable.

A byproduct of all this is any catalog containing individual stars/planets/objects from another galaxy or cluster is going to have to start with a header saying so.
Lacking that header defaults back to it belonging to the milkyway.


Janus.

[Lafuente_Astronomy]

Actually, that makes sense as a whole. I agree with the statement you said that it has to be a bunch of catalogs, like a set of encyclopedias. However, in reference to the Universal Identifiers, they must all have membership in that catalog, the specifications can follow through from there.

So, the way I see it, I see it the same way you do, just that I see it as a very immense collection of encyclopedias, each book with thousands upon thousands worth of pages. The Letter "A" of that Encyclopedia alone would have numerous books, as each focus on the different combinations of the letter "A", so to speak. The Universal Identifier then would be represented by the fact that this immense encyclopedia is created by a single publisher. Simply that.

A byproduct of all this is any catalog containing individual stars/planets/objects from another galaxy or cluster is going to have to start with a header saying so.
Lacking that header defaults back to it belonging to the milkyway.

Exactly the point. Going back to my example, if Identifier 0 is The Milky Way, then Identifier 1 is the nearest galaxy to the Milky Way, and so on. Of course in the long run, there'll probably be a more complicated but more logical, and overall better way of cataloging objects in Celestia and giving them headers. I'm sure the devs will find a way to make and manage all those headers, and I'll wait and see what they'll cook up.

Added after 7 minutes 49 seconds:
And besides, headers are essential in identification. Absolutely essential.

[Janus]

Something I just stumbled over that others might find amusing.

http://www.astronexus.com/hyg

https://github.com/astronexus/HYG-Database

While it won't solve any problems here directly, it does provide an existing reference point to compare.
It shows stars with multiple identifiers and lots of data.

I will be looking the data over as I have time.
There is this presentation about high frequency polling in an interrupt driven system I get to give.
If that is, management can correlate door knocks to polling, and email/text messages to interrupts.


Janus.

[Lafuente_Astronomy]

Looks great. It's specific enough to include information from other catalogs. But I'm certain Celestia's catalog will be a larger and more ambitious version of that. That being said, the way the catalog orders the information is good. But because this is Celestia, it has some fields that already measure the stars themselves. However, in light of more specific updates and new catalogs providing newer, more accurate and more specific data than ever before, I suggest that for an upgrade to Celestia's Star information text field, it should go like this:

-Universal Object Identifier Designation (If the developers will put a universal database catalog in Celestia)
-Main proper name (If it has one)
-Secondary proper names (If the star has any secondary proper names)
-Bayer Designation (If it has one)
-Flamsteed Designation (If it has one)
-GCVS Designation (If it has one, and if it's a variable star)
-Astronomical Catalog Identifiers/Designations (If the star is a member of numerous catalogs)
-Constellation which contains it
-Distance from The User
-Distance from Earth
-Estimated Age of the Star
-Absolute(Apparent) Magnitude
-Star Type
-Star Class Spectrum Type
-Star Class Color Index
-Apparent Diameter
-Surface Temperature
-Solar Radius/Diameter
-Rotation Period
-Number of Planetary Bodies
-Number of Natural Satellites and Artificial Satellites (Natural Satellites/Artificial Satellites)
-Number of Minor Planets, Asteroids, Planetoids and Comets (Minor Planets/Asteroids/Planetoids/Comets)
-Number of Other SSO (Solar System Objects)
-Right Ascension (Equatorial, Ecliptic, Azimuthal/Based on The User's Location, Galactic, Supergalactic/Universal)
-Declination (Equatorial, Ecliptic, Azimuthal/Based on The User's Location, Galactic, Supergalactic/Universal)
-Star Coordinates (Equatorial, Ecliptic, Azimuthal/Based on The User's Location, Galactic, Supergalactic/Universal)
-Stellar Parallax (Equatorial, Ecliptic, Azimuthal/Based on The User's Location, Galactic, Supergalactic/Universal)
-Proper Motion By Axes (Equatorial, Ecliptic, Azimuthal/Based on The User's Location, Galactic, Supergalactic/Universal)
-Proper Motion Position Angle (Equatorial, Ecliptic, Azimuthal/Based on The User's Location, Galactic, Supergalactic/Universal)
-Proper Motion Speed (Equatorial, Ecliptic, Azimuthal/Based on The User's Location, Galactic, Supergalactic/Universal)
-Mean Sidereal Time (Equatorial, Ecliptic, Azimuthal/Based on The User's Location, Galactic, Supergalactic/Universal)
-Apparent Sidereal Time (Equatorial, Ecliptic, Azimuthal/Based on The User's Location, Galactic, Supergalactic/Universal)
-Time of Rise (Equatorial, Ecliptic, Azimuthal/Based on The User's Location, Galactic, Supergalactic/Universal)
-Time of Transit (Equatorial, Ecliptic, Azimuthal/Based on The User's Location, Galactic, Supergalactic/Universal)
-Time of Set (Equatorial, Ecliptic, Azimuthal/Based on The User's Location, Galactic, Supergalactic/Universal)
-Short Description on the Star (Optional, could only be activated if "Verbose" is clicked)


I based this on some of the fields in the Database found on the Astro Nexus HYG Catalog, as well as a few specific fields from Stellarium. To the developers, you are free to include this or not. In fact, I think you guys absolutely have better judgment than me on what fields to include in the star information text field, and information text fields in general of future Celestia updates. In short, I only suggested this. The action to include all these, some of these or not is entirely up to you.

[Janus]

After looking the info in hyg over, I decided it looks close enough for some of my experiments.
While I am going to check and recheck some of the math to be sure of it of course.
It provides both position in xyz, and movement in dxdydz {delta xyz}, which I have been working on calculating.
The answers give enough feedback and accuracy to act as a guide.

I am going to make a personal fork with a custom database that includes movement.
Then I am hoping to make star positions into lines, which can have zero length.
Once I have done that, which is not going to happen quick since I am still stumbling around with opengl.
I will introduce celx commands to set star start date, and star end date.
Those dates will represent the beginning and ending of the star position lines.

Thus I can script a crawl back of star positions to the past, or a leap forward.
As long as the dates stay with in a meg year or two, galactic rotation shouldn't be a real issue.
Which will necessitate a list of star deaths, after all, it won't do to show a neighbor after it passes away.

Notes whispered into the void.


Janus.

[Lafuente_Astronomy]

Thus I can script a crawl back of star positions to the past, or a leap forward.
As long as the dates stay with in a meg year or two, galactic rotation shouldn't be a real issue.
Which will necessitate a list of star deaths, after all, it won't do to show a neighbor after it passes away.

At an estimate, how many years can you allow for star movement? Knowing that stars move really fast but take a really long time, perhaps some hundreds of millions of years will do.

[Janus]

Orbiting the milkyway takes ~230My, which means our sun is closing in on its 21st birthday, in galactic terms.
Or somewhat over 1.5Degrees per MY {mega year}.
If we stick to about 10% drift error, close enough for eyeball, but not scientific measurement.
That gives us just under six degrees of arc before it distorts visibly.
About plus or minus 3Degrees, which is ~2MY plus or minus.
{Drift error approximated by sin(0.1)^-1 = ~5.739Degrees, and I could be wrong, but we are talking about chords & secants.}

So somewhere around 4MY of displayable drift, or further if you are willing to live with distortion.
Going beyond that will require computing not only our orbit, but all the other stars as well.
It will also mean converting to a galactic standard positioning system, then calculating our view from that.
We would need to use extreme deep space objects as a clock pointer.
Andromeda and our other immediate neighbors would move quite a lot in a mere 65MY if you were wondering what stars the triceratops that died in the KT event last looked at, and so would not be useful as a clock.
Something at least a GY out as close to being on the galactic ecliptic as possible would be needed.
That would also mean detecting the galactic ecliptic from the orbit of the probable twin black holes at the center.
I am not completely sure of that last, but it seems likely to me, though I have been wrong before.


Janus.

[Lafuente_Astronomy]

Alright. That makes sense. After all, we suggested that there be proper motion in Celestia as well, starting from the earliest possible year Celestia can start at.

If this is done, we can also see not just how the stars change position but also how the position of "Brighest Star in the Sky" is actually a title that gets constantly transfered from one star to another. 5 Million years ago, Adhara(Epsilon Canis Majoris) was only around 34 LY from the Sun. But because it's an immensely large and bright star, it was the brightest star in the Sky in that time, and was even brighter than Venus

That being said, if all star's movements were properly measured(Hopefully GAIA does that), who knows which stars will come into the Earth's life and illuminate it as the arrangement of stars in the sky changes.

[Janus]

The north star changes over time.
Closest star changes over time.
Brightest star changes over time.

I am curious to see what we once saw, as we saw it.
Such as these which are little redundant.

https://phys.org/news/2018-11-prehistoric-cave-art-reveals-ancient.html
https://www.breakingnews.ie/world/ancient-animal- ... tar-constellations-888247.html
https://www.businessinsider.com/ancient-cave-draw ... onstellations-of-stars-2018-12

An argument in anthropology is about animal drawings.
In part, I wonder if some of the star signs were ones that corresponded to the best time to hunt the animal it represented.
I also want to see what the signs looked like back when they were first settled on.
Are cave paintings stylized figures for animals, or reproductions of constellations.
Perhaps the animal in a group with stars in it, represented the time of year the hunt took place.

How closely did the pyramids at Giza match Orion's belt back when they were built?
Or is that correlation some sort of modern thing since they more or less match now?
Has halley's comet, which I theorize/suspect is locked in resonance with Venus/Jupiter/Orcus, ever shown up during a lunar or solar eclipse?
I am filled with questions, and this program is just one way I work on finding answers.


Janus.