While I love the ASCII art, yeah... thats not how the moon phase works. At half moon it should be exactly half occluded, so like it was cut in half, not just a weird circle cut out.
This is something that partially bothered me about GTA San Andreas on Ps2. They used the same trick as here but I would give them a pass because it was just a small detail on a large simulation on an already very limited system.
There are some great bones here on the site, just need to tweak the shadow and it can be a 10 out of 10.
Yeah they seem to have drawn it as a dark circle going across for some reason.
But earth's shadows aren't what create moon phases IRL. The real moon phases are from the sun lighting the moon from different angles so this is just super weird.
yet an incredibly basic and stupid oversight when the only point of your site is to show the phase of the moon.
I'm voting that this was vibe-coded by an LLM on behalf of someone who not only didn't write the code but didn't bother to look at a picture of the moon or look at the sky before deploying it. If so, it's almost the perfect Platonic example of "what could possibly go wrong?"
This is also the standard way to present moon phases in watch complications:
> One of the objections raised to a conventional moonphase display, which shows the visible part of the Moon via an aperture in the dial, is that it is not an accurate representation of what you see when you actually look at the Moon over the course of a month.
Well, let's hope nobody drills a hole in the Moon (or blows the Moon up) in the next couple thousands of years, for the sake of those watches' usability.
As a classic Neal Stephenson (great world building, good plot... and it ends)... Seveneves starts with:
> THE MOON BLEW UP WITHOUT WARNING AND FOR NO APPARENT reason. It was waxing, only one day short of full. The time was 05:03:12 UTC. Later it would be designated A+0.0.0, or simply Zero.
... and given the book and the "what happens"... a watch surviving (much less keeping the phase of the moon) would be impressive.
It’s actually safe to do that, watch owners would be distracted by global cataclysms ;)
> Tides would be much smaller (…) But the movement of tides underpins the balance of ecosystems the world over, and an impact that widespread would cause global biological collapse across oceans and, in turn, the whole Earth.
It's amusing that it isn't truly a circle. It's a box with rounded corners and the corner rounding radius chosen to be half the box width. There is probably not a simple CSS trick like that for doing the correct 2D projection of a hemisphere.
That's a great suggestion for a more accurate render, but I want the author to know that I still think they did an amazing job. It's fun and it looks great. (Your feedback was really constructive, and I don't think the author will misinterpret it. Some of the other commenters here, on the other hand...)
I feel like ascii art loses something when it's not sized to a standard text mode (at least to width, like 132x132 is fine). At some point you're just using weird pixels and this is approaching that for me. Same goes for changing the color of the characters continuously; terminal colors are cool. I'm probably just a crazy purist
It's not even just that: the shadows go mid-character, instead of using characters as pixels. It is just not ASCII ART at all, just some ASCII characters used as a filler.
oooh. Good point. I didn't even noticd that at first glance as I was floored that anyone would show off something so totally wrong. But yes, not responsive ascii art if you're just throwing a transparent circle over the same bunch of characters.
And it's within a circular mask. Half the challenge of ASCII art is creating the illusion of smooth curves and edges when you mostly just have blocky letters. There's barely any attempt at smoothing the edges at all.
I actually feel like using color at all is cheating. I'm okay with a wider width than 132 px, but it needs to be sized such that you can clearly make out the letters, which I can't in the TFA.
Yeah, this rendering is Just Wrong - moon phase is viewing a half-lit sphere from different angles, no external shadows are involved. Maybe look at https://github.com/chubin/pyphoon instead, which claims a history including Jef Poskanzer code going back to the 1970s...
> ASCIIMoon is a small web app that tracks the moon's phases and uses ASCII art to create a basic visual representation of the moon's current appearance based on light percentage.
> This is a personal project, and is in no way a precise representation.
https://www.moongiant.com/phase/today/ is what it should look like. The moon for July 2nd should be something that is a half circle illuminated and a half circle in the dark.
The image on asciimoon shows a circle occluding the moon where it appears that the circle occluding the moon's edge is at 50%.
I'd have to sit down and do the math, but there is way more than 51.79% illuminated in today's rendering.
While I recognize that this is a personal project and not a precise representation... it has a fair amount of work to do to make it so that the correct percentage is illuminated.
This does have some interesting JavaScript and css tricks... but it needs some more math done.
It seems misleading to say that this is "rendered in ASCII art"... more like the phase is being rendered, and it has an ASCII art moon background. If I copy and paste the moon text into a text editor, the phase is gone, so it's not actually part of the text.
And as others have pointed out, phases don't work like that.
This looks really nice (apart from the shape of the gibbous moon, but others have already pointed that out).
Could you possibly detect the user's approximate location and rotate the whole thing 180 degrees if they're in the southern hemisphere? Down here, the moon looks technically the same, but it's 180 degrees* rotated because we're standing upside-down when we look at it. The craters are flipped and the light comes from the opposite side.
*Actually it's only rotated a full 180 degrees at opposite poles, and the exact rotation depends on your latitude. But perfect is the enemy of good etc.
The match between the ASCII part and the clipping circle breaks when zooming out too much on a desktop browser.
Also, given sufficient character "resolution", ASCII art approximates pixel art. This isn't very far from it, with (on my screen) characters of 3x7 pixels. And would require a 200x78 character terminal to fully display.
The astronomical criticism here is warranted. But where are the usual defenders of "art" here to take arms for this project's interpretation of "ASCII Art"?
It looks cool and I imagine it wasn't easy to do. And the color of the moon changes when you reload the page.
I remember going to a planetarium as a kid and seeing a booth where you could “buy land” on the Moon (or any other planet) and get a fancy certificate to prove it
This project totally brings back that kind of charm, but with a digital twist.
I want the shadow that progresses across the moon to match reality. A crescent shape that always goes through the north and south poles — that flattens as it approaches a quarter moon....
we do we need a web site for an ASCII art display? I just want a command I can locally can in my terminal. And if I want it online, just a telnet server ;-)
First, there comes a point when it’s not ascii art, it’s just dithering. The use of different colors for the characters goes even further from ascii art.
Second, opening this on my iPad results in a moon with a black crescent on the bottom and an oddly shaped dark green crescent on the left. What star system is this moon in?
This is something that partially bothered me about GTA San Andreas on Ps2. They used the same trick as here but I would give them a pass because it was just a small detail on a large simulation on an already very limited system.
There are some great bones here on the site, just need to tweak the shadow and it can be a 10 out of 10.
But earth's shadows aren't what create moon phases IRL. The real moon phases are from the sun lighting the moon from different angles so this is just super weird.
I'm voting that this was vibe-coded by an LLM on behalf of someone who not only didn't write the code but didn't bother to look at a picture of the moon or look at the sky before deploying it. If so, it's almost the perfect Platonic example of "what could possibly go wrong?"
(for example the comments are all capitalized, there's inconsistent indentation)
Assume y vertical and x horizontal and z out of the page.
The moon is a disk. x^2 + y^2 <= R^2
For all the points, calculate z = sqrt(R^2 - y^2 - x^2)
Let the vector v(theta) = <sin(theta), 0, cos(theta)> point to the Sun.
Points <x,y,z> • v(theta) >= 0 should be bright, The rest dim.
* https://github.com/Sean-93/asciimoon/blob/main/src/component...
> One of the objections raised to a conventional moonphase display, which shows the visible part of the Moon via an aperture in the dial, is that it is not an accurate representation of what you see when you actually look at the Moon over the course of a month.
https://www.hodinkee.com/articles/the-beautifully-pointless-...
https://what-if.xkcd.com/46/ - while it deals more with the Earth (which has a more molten core), its still applicable. There's still molten material ( https://science.nasa.gov/moon/composition/ ) but it's a lot deeper.
> (or blows the Moon up)
As a classic Neal Stephenson (great world building, good plot... and it ends)... Seveneves starts with:
> THE MOON BLEW UP WITHOUT WARNING AND FOR NO APPARENT reason. It was waxing, only one day short of full. The time was 05:03:12 UTC. Later it would be designated A+0.0.0, or simply Zero.
... and given the book and the "what happens"... a watch surviving (much less keeping the phase of the moon) would be impressive.
There's also the Dr. Who take ("Spoilers") https://youtu.be/pHOnGSFzd3Y
> Tides would be much smaller (…) But the movement of tides underpins the balance of ecosystems the world over, and an impact that widespread would cause global biological collapse across oceans and, in turn, the whole Earth.
https://www.popularmechanics.com/space/moon-mars/a43633761/b...
>To address this problem, moonphase displays were invented which use a spherical miniature Moon
I have two additional feature requests:
1. Persistent URLs (replaceState)
2. Lunar eclipses
https://www.roysac.com/tutorial/rowanasciiarttutorial.html
It's ASCII art. Drop the bitmap mask.
This looks more like what is seen during a solar eclipse.
> ASCIIMoon is a small web app that tracks the moon's phases and uses ASCII art to create a basic visual representation of the moon's current appearance based on light percentage.
> This is a personal project, and is in no way a precise representation.
https://www.moongiant.com/phase/today/ is what it should look like. The moon for July 2nd should be something that is a half circle illuminated and a half circle in the dark.
The image on asciimoon shows a circle occluding the moon where it appears that the circle occluding the moon's edge is at 50%.
I'd have to sit down and do the math, but there is way more than 51.79% illuminated in today's rendering.
While I recognize that this is a personal project and not a precise representation... it has a fair amount of work to do to make it so that the correct percentage is illuminated.
This does have some interesting JavaScript and css tricks... but it needs some more math done.
And as others have pointed out, phases don't work like that.
At least there are no stars in the shadow, so they get partial credit.
https://en.wikipedia.org/wiki/Star_and_crescent#Contemporary...
https://slatestarcodex.com/2019/05/02/little-known-types-of-...
I think this is my most popular repo on github.
Could you possibly detect the user's approximate location and rotate the whole thing 180 degrees if they're in the southern hemisphere? Down here, the moon looks technically the same, but it's 180 degrees* rotated because we're standing upside-down when we look at it. The craters are flipped and the light comes from the opposite side.
*Actually it's only rotated a full 180 degrees at opposite poles, and the exact rotation depends on your latitude. But perfect is the enemy of good etc.
We are standing the right way up your northernist simp.
They are the ones who are wrong. Never forget it.
Also, given sufficient character "resolution", ASCII art approximates pixel art. This isn't very far from it, with (on my screen) characters of 3x7 pixels. And would require a 200x78 character terminal to fully display.
It looks cool and I imagine it wasn't easy to do. And the color of the moon changes when you reload the page.
It took me straight back to childhood.
I want the shadow that progresses across the moon to match reality. A crescent shape that always goes through the north and south poles — that flattens as it approaches a quarter moon....
First, there comes a point when it’s not ascii art, it’s just dithering. The use of different colors for the characters goes even further from ascii art.
Second, opening this on my iPad results in a moon with a black crescent on the bottom and an oddly shaped dark green crescent on the left. What star system is this moon in?
https://www.acme.com/software/phoon/
You can also use xphoon(6) on Linux/*BSD to show a kind of ASCII moon on your root window.
https://formulae.brew.sh/formula/phoon
curl https://wttr.in/Moon