Episode: 2598
Title: HPR2598: Calculating planetary orbits in Haskell
Source: https://hub.hackerpublicradio.org/ccdn.php?filename=/eps/hpr2598/hpr2598.mp3
Transcribed: 2025-10-19 06:22:19

---

This is HPR Episode 2598 entitled Calculating Planetary Orbit Sympathical.
It is hosted by Tutoto and is about 29 minutes long and currently in a clean flag.
The summary is Tutoto talks about Calculating Planetary Orbits.
This episode of HPR is brought to you by AnanasToast.com.
Get 15% discount on all shared hosting with the offer code HPR15, that's HPR15.
Better web hosting that's honest and fair at AnanasToast.com.
Hello and welcome to the HPR.com. I'm your host Tutoto and this show will be a
part of HPR15.com.
The program I'm talking about here is based on the guide by Paul Schleider.
This link to his webpage in the show notes.
I recommend you to check that out because that has a good explanation of these things.
It's also a bit more atoling on the show notes on the Wikipedia about the terms that
I'm going to use.
I'll try to explain them here, but the Wikipedia articles are probably a lot better.
So first, build about coordinate systems.
In general, fixing a location of an object requires a reference point or a regga.
And since everything in the space is moving, there's many, many different coordinate systems
to pick from.
And in a given problem, a correct sort of coordinate system makes calculations health easier,
for example.
And nothing prevent you from switching from one system to another and as the need arises.
For example, if we were to travel from Earth to Mars, we probably would start from the
Geosynthetic system, the system may be the only place around the Earth.
And at some point, when we leave the Earth, switch to the Heliosynthetic system, the center
is the Sun and planets are moving around it.
And when we enter into the orbit of the Mars, we would switch to the Mars center system
at the Mars center.
Because less than moving things, rather funny moving things, the easiest is to do the calculations.
And this does not even take account that you have to move on the surface of the planet.
You have yet another set of coordinate systems.
Next about orbits that we are going to calculate.
So we are going to calculate how the planets are moving in the space.
But the orbits of the planets are elliptical and they are slightly tilted and they change
slowly over time.
And there are six orbital elements that can be used to define shape of the orbit in
location of the planet on that orbit.
And the first one of those is the X and X in the recipe, as I said, orbits are elliptical
and X and the recipe is just how elliptical they are, how they are completely circular
and they are really very squished.
And if the L-L-R, if X and the recipe is high enough, that means that the orbit is in
the elliptical anymore, it's a parapoloid, parapoloid, meaning that the objects are
just swings past, for example, some, and then it comes back.
Next one is semi-mature axis.
This is a sum of periapsis and apoapsis divided by two.
And periapsis and apoapsis are the points on the orbit that are closest and farthest away
of the body that we are orbiting around.
And this, basically it's the, if you take the elliptical, the longest possible one you
can inside of it and half it is the semi-mature axis.
Next one is inclination, that tells how till that the orbit is, because the orbits aren't
completely on the level, interrequact to the restaurant's plane, I mean, they could be
but they usually are, this tells, tells how much, how many, how many decreases till that,
which means that when there's an inclination, the planet is sometimes over the reference
plane and sometimes under the reference plane.
And a reference plane in this case, when we are talking about the planets orbiting the
Sun, is the elliptical plane, which is defined by the plane that the Earth is rotating around
the Sun.
So actually the Earth has a zero inclination.
Next one is longitude of the Asken node, this tells where the orbit transfers from below
to above of the elliptical plane, because not all the ellices aren't.
And in the same direction, with the long tail pointing to the same direction, this defines
where exactly, if the ellipses are rotated.
And last one is the Asken node last one, next one is the argument of periapsis, sometimes
this is called the argument of perihillion, or argument of peri-k, dividing on what you're
orbiting, but periapsis is the general term peri-k is when you're talking about objects
orbiting around the Earth and perihillion is when you're talking about objects orbiting around
the Sun.
But this one defines the rotation of the ellipse on the orbital plane, like you can imagine
this by sticking the stick on your center of the orbit and rotating it slightly, this
rotates the ellipse on the plane that is defined by the orbit itself.
Last one is the true anomaly.
This defines the position of the planet as an annual orbit and some given special time
called epoch.
And I have linked a Wikipedia article that has a nice little drawing about all these elements.
I recommend you to check that out because it's a lot easier to see the picture and understand
how good this is, how the angles and axes relate to each other.
It's rather difficult to explain it while speaking.
And because we're dealing with the computers, we're dealing with the radians.
We have a trigonometric function that has to use radians.
I think you need of the angle.
So I had to write a couple of helper functions to convert the two and from radians.
I'm basically between the decrease and radians and then there's a little helper function
that clamps the decrease into 0 to 360 degrees.
That's not really...
Well, that isn't really needed, but it makes a output of the program slightly nicer
because it's easier, at least for me it's easier to think that there's a handle of 10 degrees
and the angle of 370 degrees.
It's a super big difference to the result of the calculations.
But in any case, I made one that was suggested by the slide anyway.
Slide, sorry.
And then, like I said, the orbital element change over time.
So we need a time.
That is expressed as a fraction of days since the 31st of December 1990.
Okay, now we know we're talking about...
I can start going through the actual calculations.
I'm not going to read out the equations or even try to explain how they do the different transformations
for that it's a good idea to check the tutorial that I linked.
So I made this as an experimental program.
I have written this in several times with various languages.
Now I wanted to try to write it in Hasker and see how the type system can be used to enforce calculations
to use the correct type.
There's a plenty of different coordinates, plenty of different angles.
And I wanted a type system that makes sure that I don't mix for example positions in different coordinates.
This is... this kind of, of course, done with pretty much any subjectly type language.
So I started by defining sound earlier than Mercury.
I just argued, I'll get write data files without any parameters.
So I'll just say that something is sound earlier than Mercury.
And then I defined a type class orbit that you see.
That has no name, such as defines orbital parameters.
This has two type parameters, sender and body.
This way I can express on a type level that Mercury is rotating sound and Earth is rotating the sound.
And if I were to try to calculate how Mercury moves in relation to the Earth,
I would mix up all those various numbers.
I would be able to calculate location in sound coordinate system of the Earth in respect of the sound.
And Mercury in respect of the sound and then translate them and get the location of Mercury in respect of the Earth.
That's why I defined this with two type parameters.
And this requires turning on the multi-param type X classes and flexible instances and flexible context extensions for the hospital.
I did not 100% sure what all those actually do.
I know that multi-param type classes is needed to be able to define path classes with two parameters.
But those later on, I had to turn on to get the program to compile.
As you might have guessed, I'm not technically good at the hospital.
It's more of a hobby of mine where I play with that now.
So in this type class, orbit has functions that are used to calculate those orbital parameters that I mentioned earlier.
With a given date, given day.
So if I were to compile longitude of ascending node and give that to two parameters, Earth and Sun,
and some date, that would give me the longitude of ascending node of Earth orbit around the Sun on that given day.
And return values of all these functions are also...
Also, I can track data lives with two parameters.
So the longitude of ascending node that is given me actually encodes that this is Earth's longitude of ascending node around...
On an orbit around the Sun, and it has this value.
Same with the inclination of eclipse argument of periaxis.
Same with major axis eccentricity and mean anomaly.
It just makes it easy to deal it in the Sun, you have a bunch of angles floating around.
And as I mentioned, Haskell uses radians, like pretty much, I suspect, almost every programming language.
And I cannot visualize radians that easily on my head.
I have a rotor, I have an algorithmic data type decrease, that can wrap anything,
and type class decreaseable, that is type class with a single parameter, and single function to deck.
So if I create an instance of...
If I create an instance of...
For example, mean Anna of that decrease, that means that any mean anomaly,
where for any orbit, I can turn into decrease.
It's just simplified saying to deck, and the value of the mean anomaly.
And it actually wraps the mean anomaly value into decrease,
algebraic data type, because this changes the type,
so I don't accidentally use this converted value somewhere laid on the program.
Anything that isn't wrapped into decrease is radians, and anything that has been wrapped to the decrease,
cannot be used in the calculations.
So now that we can calculate our orbital elements, in respect of give and date,
because they slowly wander around, we can start working on the position on the orbit.
And this is first, we need the eccentric anomaly.
This is why to research what this actually is, but I couldn't figure it out.
Then I'll test the three major...
There's three anomalies, eccentric anomaly, mean anomaly, and true anomaly,
that all relative to the angles on the orbit.
And the true anomaly is the location of the planet as an angle.
And the mean anomaly is a sort of mean value.
Like, if the planet were rotating an orbit in a completely circle out plane,
the whole asteroid orbit, then the mean anomaly, and true anomaly,
would give the same answer, I think, you better understand.
But anyway, first we need to calculate this.
We have a mean anomaly.
This is convenient mathematical angle, I think, to all.
But we need to turn that into a true anomaly,
so that we know exactly where the planet is on the orbit.
And we do that by first calculating the eccentric anomaly, anomaly,
which needs the mean anomaly and the eccentricity,
how is this the orbit is, and for example, the eccentric anomaly,
and with this, we can combine with the eccentricity again, get the true anomaly.
So this gives you the angle of the planet is on the orbit.
And then we calculate the distance of the object.
The object we are orbiting around, and the object that is doing the orbiting,
which uses the eccentric anomaly, eccentricity, and semi-major axis,
and produces distance between these two objects.
Again, distance is an algebraic data type,
so it actually states that this is a distance of, for example, sound and memory.
It is pretty convenient.
Yeah, it's a pretty convenient.
And with this two numbers, we have to fix the relative location of the object.
Of course, this is in a polar.
No way.
I think it's a polar coordinate system.
And it's on a orbital plane.
So we know the distance, and we know the angle between our reference point
and the angle that the planet is currently.
But this is, like I said, it's an orbital plane.
It's a plane that is defined by the...
Basically, basically, it doesn't take the inclination into the account.
And for this, we want to rotate our coordinate into ecliptic coordinates.
This is the coordinate system that is defined.
For example, with sound in the orega, and with a three-dimensional grid.
You come on x-watched coordinates.
And if I can move correctly, the set axis points to the upwards of,
from the ecliptic airplane, as much as the upwards in the space.
I think it's a node.
I think it's defined that the side that has the node, both of the Earth, is the up.
And the x-points do the vernal equinox.
But that's not so important when it comes to this episode.
We just want to rotate into a three-dimensional coordinate called ecliptic coordinates.
And from here, we need the Ruanomali distance,
because this is the location of the planet that we are.
The location is what we are rotating.
Longitude of ascending node, which...
And argument of periaxis and inclination to ecliptic.
And with a bit of three-conometry, we can rotate our orbit
or our location on the ecliptic coordinates.
And if we want to calculate where the stuff is,
when viewed from the Earth, this is not yet enough.
Of course, there was this information to the Sun.
If we were standing on the surface of the Sun, then we could use this.
Continued using these coordinates, but we are not standing there.
So we need to translate there.
We need to transfer the order equal to our coordinate system from the Sun to the Earth.
And this we do by calculating the location of the Earth.
Actually, the tutorial uses the system later that treats Sun as orbiting around the Earth.
But it's just a matter of point of view, basically.
But we calculate the location of the Earth.
We have the location of the aadat planet, Mercury in our example,
and then we just end the boat of the Moon using the same coordinate system
as we just subtract those coordinates from each other.
And we end up with the Mercury's location,
on the Earth's center of the ecliptic coordinate system.
This is starting to get where we want.
But this is not all yet.
The three-dimensional coordinates are not that easy to issue a point from the Earth
or are they contain extra information we don't need.
So we are moving things to the equatorial coordinate system.
And this you can imagine that you will slice the Earth from the equator.
From the latitude of zero degrees.
And the plane that is defined by that is the equatorial plane.
And on this coordinate system we express things in a right-ascension declination.
The right-ascension tells the direction on the plane.
And it's given in hours, actually.
Because the Earth rotates around in 24 hours.
So the plane is divided to 24 hours,
and those are divided into the minutes and seconds.
And declination is the how angle of the object from that equatorial plane.
So zero degrees is on the plane and 90 degrees is at the zenith.
So somewhere really close to the polaris.
And for this we need the ecliptic coordinate,
coordinates given in respect of the Earth, not on respect to the Sun,
and currently we can calculate the location of the planet in the equatorial coordinate system.
And this is now at the point where you can go outside and point to the sky and say that,
okay, my computer program says that that pride dot all the day is Mercury.
So this concludes our calculations for now.
There's a future improvement that could be added,
like you could translate into the horizontal coordinate system,
which takes into account your location on the surface of the Earth.
Because we are not standing in the middle of the Earth.
We are standing on the surface of the Earth.
So we will need to rotate things yet again and take day of the time of the day into account.
So we could get our angles in respect of the horizon of the local position.
This would make it even easier to find whether planet is on the sky.
And another thing I might add later is to calculations for perturbations.
Because these equations are used so far.
I assume that there's only two objects.
The thing that we are opening around and then the thing that we are opening in.
So they are nice and clean and easy calculations.
Very easy when you copy them from the net.
But there's nine planets in the solar system.
All of them are packing each other constantly.
All of them are opening.
There's some huge translate Jupiter and Saturn.
So every time those planets swing by,
they are messing up with the opposite of another planet.
So the opposite, the opposite wobble somewhat.
So we would have to...
It's not too difficult.
It's not too difficult.
But I haven't written that part yet.
But we would have to take into account how those big planets change the opening of another planet to get into even that accuracy in these fault places.
These are really enough that you can really go outside and point at.
That's the pride of this.
This is the planet I'm interested in.
But that concludes this twist of episodes.
If you have any questions or comments,
please let me know or even better record your own episode.
See you later.
You've been listening to Heka Public Radio at HekaPublicRadio.org.
We are a community podcast network that releases shows every weekday, Monday through Friday.
Today's show, like all our shows, was contributed by an HPR listener like yourself.
If you ever thought of recording a podcast, then click on our contributing to find out how easy it really is.
Heka Public Radio was founded by the digital dog pound and the Infonomicon Computer Club,
and is part of the binary revolution at binrev.com.
If you have comments on today's show, please email the host directly,
leave a comment on the website or record a follow-up episode yourself.
Unless otherwise stated, today's show is released on the Creative Commons,
Attribution, ShareLite, 3.0 license.