corbeau217/mixed_orrery

solar system model done in whichever language i feel like it should be in

mixed language orrery

---

quicklinks

• ~/readme/About
• ~/readme/Status
• ~/readme/References

---

About

• repo is for making solar system models done in whichever language i feel like it should be in

show more

• this will inherit from the work done in
  • scala_drawing
  • scala_ui_testing
  • cpp_isometric_grider
  • ProceduralCpp
  • various other private repositories
  • various content done for university assignments
• aim is to keep things as simple as possible to create the base structure which can then be made into something more.
• as it stands (at repository creation), both starting points have already suffered massive feature creep and will be simplified to just:
  • 1 star at centre
  • planets orbit star
  • moons orbit planets
• ships/asteroids will be labeled as satellites and will be a milestone for each implementation
• everything should be kept as whichever portable drawing system is on hand easily for the given language that requires minimal starting time.
  • further modifications for UI and other features can be done once foundation is setup
  • changing drawing systems can then be treated as a new sub-project with the same language
  • care should be taken to avoid initial use of any system that requires pixel based drawing until this project has more implementations and familiarity
  • for the above point reason, ncurses should be secondary to raylib/raygui unless we find a function that handles circles or can use quick math to do it
• for inital plans, we'll probably stick to the Kerbol System from Kerbal Space Program 1 as it's what is most familiar in mind right now
  • inspiration didnt necessarily come from this, as it has been a long time that I've wanted to create this project.
  • watching For All Mankind and playing a lot of Kerbal space program 1 lately has meant that my interest was returned and so could dust off old notes and plans

---

Implementations

Status

• 
  show/hide details

  • [12/02/2024] - alpha with some feature creep, check todo
  • [13/02/2024] - somewhat prototype
  • show/hide todo

    • [12/02/2024] - need to fill in the other planets and moons
    • [13/02/2024] - changing over structure to list of orbiting satellites on a body, and those orbits hold what the thing is
    • [13/02/2024] - move stage paint code to solar system painter
    • [13/02/2024] - restructure window/app to be seperated out from main
    • [13/02/2024] - restructure the code to allow keyboard input
    • [13/02/2024] - restructure the code to allow mouse input
    • [13/02/2024] - add in wasd moves system center offset around
    • [13/02/2024] - add in q and e changes the zoom level of the sim
    • [13/02/2024] - change over to BigDecimal
    • [13/02/2024] - status text stuff
    • [13/02/2024] - details of placement of things showing (red/yellow/green based on percentage of max)
    • [13/02/2024] - break circle into points of known values and work between them
    • [13/02/2024] - need labels of which planet is which
    • [13/02/2024] - need status of time speed
    • [13/02/2024] - need status of scale
    • [13/02/2024] - wikis have a decent explanation of orbits too
    • [14/02/2024] - change to use unit circle points and multi with value??
    • [14/02/2024] - placeholder?
  • show/hide changes

    • [13/02/2024] - changed structure to having orbit list and each orbit holds ref to body
    • [13/02/2024] - changed to name of universe to solar system
    • [13/02/2024] - made the solar system painter class
    • [13/02/2024] - added in stubs for keyboard input
    • [13/02/2024] - added in code for updating the phase of each orbital
    • [13/02/2024] - added in code for centering on a body by reference
    • [13/02/2024] - added screenshots to readme
    • [13/02/2024] - wasd controls scales
    • [13/02/2024] - f cycles planet/sun focus
    • [13/02/2024] - added more bodies to the system
    • [13/02/2024] - b toggles larger/smaller planet scale
    • [13/02/2024] - changed scaling over in an attempt to reduce precision loss, but will need to change to BigDecimal to solve it
    • [13/02/2024] - placeholder?
  12/2/2024 - show/hide screenshots

  
  12/2/2024 - show/hide screenshots

  
• more to come

---

References

• Orrery
• solar system model
• Kerbol system
• solar system
• 3blue1brown linear transformations
  • rotation of vectors
• pi
• double precision
• unit circle explainer
• unit circle rolling toy

---

brainstorming

brainstorming pre-java sub-project

• original typed brainstorm [8:30AM][11/02/2024]

  show / hide

  Have the main Just does the applet code jpanel/jframe/paint loop, standard stuff

  Then you have stage, has the universe scale and a reference for sun

  The sun is a celestialBody which has things orbiting it Those orbits are celestialBodies or ships They have semimajor axis, eccentricities and stuff Then you have the info about the body

  Could seperate out the data from the drawn Celestial body is now obj with its details and orbit info Double linked? Or hands off data every draw?

  Maybe it doesnt know about parent so you paint. From god view and can see the parent while painting

  Painting is a depth first or bredth first? Probably bredth first

  You have the objects and orbit data You set that up and its just a store

  Then you can paint using that information based on current time to find where they are now

  Maybe paint is dumb and doesnt calculate at all. Just paints

  Now you could thread? Or you do update before a paint cycle. Theyre seperate accesses from main/paint loop level

  tick(long chronoIntervalMillis) which hands off all children as a “time elapsed” since last update

  Implication is u have a repaintTimeSince and updateTimeSince Paint is dumb and main worries about that But update figures out how long it takes and then predicts if it can do another before the next paint cycle?

  Maybe update only happens when we need to paint. So we just keep track of

  microverseOriginTimestamp lastPhysicsTimestamp lastPaintTimestamp

  If it there’s enough time since we last painted, we do a paint, but we need to ask for new context to paint so we’re not painting old content. Check buffer

  What if feature creep a buffer for paintables. So the buffer is updated with pixels of what we’re painting

  https://en.wikipedia.org/wiki/Astronomical_object

  Namables

  Star is root

  Planet is can have moons

  Moon is can have asteroids and ships

  Asteroid is can have ships

  Ships is can have nothing. Terminal point

  Star can have moon/asteroid/ship

  Maybe we categorise by size??

  Can only have object with a classification strictly less than the parent

  Becomes min heap kinda?? Maybe make rules list

  Orbitals are strictly less than or equal to the parental body

  Floats only when it’s strictly finite small distances otherwise it’s double

  Like drawing to screen could be float

  But distances are doubles

  satellites instead of ships/asteroids

  So structure

  Main Applet Stage Universe AstronomicalBody semimajoraxis orbitalperiod eccentricity angleofinclination angleofmajoraxis? Optional<List> orbitals

• rounding error brainstorming [4:00PM][14/02/2024]

  try for define the different points try for the roation of the plane with the things then matrix math try for the points around the circle returner? try for the get via multi methods then average?? we should polynomial it then we can figure out which things are adjacent/equivalent then switch between the forms like sliders between universe but it's sliding between the different types of things we can use until we're using one that's good break everything into smaller sub problems then build it back up break into matrices of smaller numbers then return for building it back up into big numbers all at once? numbers that refactor themselves in the unit circle roller, it's the yello one that's spooky the others there's 3/4 things lining up, but the yellow it's the one time it's kinda alone aside from blue 1pi/12 * 15d 5pi/12 * 75d 7pi/12 * 105d 11pi/12 * 165d 13pi/12 * 195d 17pi/12 * 255d 19pi/12 * 285d 23pi/12 * 345d oh my god we found the 8 it's the Math.sin/Math.cos lerping

  actually, maybe it's the drawn circle that's a spooker 🥲

  oh my god it is, we checked documentation we gotta make a hemkn circle drawing now 🥲

  ---

  Corbeau — Today at 16:21 ill have a better accuracy than [incorrectly targeted slander] ye idk if video shows it but it's like a sine wave in the deviation of the orbit like it's orbiting the path it's meant to take Mungo_ — Today at 16:22 what you could do is a variant of the circlePoints thing this is maybe a bit complicated Corbeau — Today at 16:23 i thought it was rounding error, bc it sounds like it would be, but it's literally just that they took the laziest approach in calculating it

  Mungo_ — Today at 16:34 so with this, you could modify the function to draw n points on an arc of a circle then it would be a matter of finding where the circle intersects your camera rectangle Corbeau — Today at 16:35 C: oooooo good point Mungo_ — Today at 16:35 this is a Fun computer graphics problem :))) aka frustum culling but not really the same as the traditional problem, because usually you just discard entire objects that aren't on the screen

  Mungo_ — Today at 16:39 wheels were made to be reinvented

  soooooo, they're all prime numbered factors of pi over 12 for the radians of each of the spooky spots??

  show / hide