Astrolabe: The Sky, Now in a Browser
Astrolabe has grown beyond its original HomeSeer plugin roots into a public web application for anyone interested in the sky. Using the same underlying astronomy engine, the web version calculates planetary and lunar positions, twilight, eclipses, occultations, satellite passes, apparent altitude, and more for any location. It uses modern astronomical models and current observational data where useful, while keeping the core calculations independent of outside services. In short, it is the same increasingly serious astronomy engine, now available to anyone with a browser.
No HomeSeer required. No subscription required. Just the sky, computed.
No HomeSeer required.
A small update to the increasingly-not-small Astrolabe project.
I have now put a public web version of Astrolabe online:
https://astrolabe.cyberlogix.net
The HomeSeer plugin is still the original project, and it remains the version that can expose astronomical information directly to automation, trigger events, and become part of the house.
But somewhere during the rather substantial accuracy overhaul, I realized something:
There really isn't anything particularly HomeSeer-specific about knowing where Jupiter is.
The astronomical engine had become substantial enough to be useful on its own.
So I gave it a web interface.
Same sky, fewer prerequisites
The web version is intended for general access.
You do not need:
HomeSeer.
A plugin.
An account.
Or any special software.
Give Astrolabe a location, either by allowing the browser to provide it or by entering latitude and longitude, and it calculates the sky for that location.
The underlying goal is the same as the plugin:
Compute the astronomy rather than simply retrieve somebody else's answer.
That distinction has become increasingly important to me as Astrolabe has evolved.
The site is not asking an online almanac:
“Where is Saturn?”
The Astrolabe server is doing the calculation.
It is the same rabbit hole
The public version benefits from all of the accuracy work I have been writing about recently.
Planetary and lunar positions come from JPL DE440.
The coordinate system uses IAU 2006 precession and IAU 2000A nutation.
Earth orientation can incorporate current IERS data.
The apparent-coordinate machinery includes the things I discovered were missing during the recent accuracy expedition:
Aberration.
Gravitational light deflection.
Proper sidereal time.
UT1.
Topocentric parallax.
Observer elevation.
And the other corrections that gradually became visible once the larger errors were removed.
So yes, the web application calculating whether the Moon is above your horizon is quietly worrying about things such as the current orientation of the Earth.
This remains somewhat more machinery than is strictly necessary to answer:
“Can I see the Moon?”
I have accepted that.
Location matters more than I appreciated when this started
One of the most interesting lessons from the accuracy work was just how important the observer's location becomes once the Moon is involved.
For distant planets and stars, moving around on the surface of Earth changes the apparent position only slightly.
The Moon is close.
Very close, astronomically speaking.
Its topocentric position can differ enormously from the geocentric one.
That means the web version is not merely showing a generic sky for “Phoenix” or “London.”
It calculates for the actual coordinates supplied.
And at the accuracy Astrolabe is now chasing, even small differences in observer location eventually become measurable.
I recently compared two GPS positions only 7.9 meters apart and found that both instruments were resolving the difference correctly.
Which led to the mildly absurd realization that walking across the property can now change the calculated lunar position by more than some of the relativistic corrections I had spent time implementing.
At some point, astronomy becomes geography.
Geometric and apparent are both shown
Another thing I wanted the web version to make clear is that there can be more than one legitimate answer to:
“How high is that object?”
Astrolabe reports both:
Geometric altitude
and:
Apparent altitude.
Geometric altitude is the result of the astronomical geometry.
Apparent altitude also includes atmospheric refraction.
Near the horizon, those can differ by roughly half a degree.
That is enormous compared with the arcseconds I have been chasing elsewhere.
When current atmospheric data is available and Internet access is permitted in the plugin, Astrolabe can use actual temperature and surface pressure to refine that refraction.
The public site also keeps the distinction visible rather than quietly presenting one number as though the atmosphere were perfectly predictable.
Because it isn't.
The mathematics may know exactly where the geometric horizon is.
The atmosphere has opinions about what you actually see.
Time belongs to the place being described
There is another detail in the web version that I particularly wanted to get right.
Times are shown in the local time of the location being calculated, not simply the timezone of the person viewing the page.
That matters if you are looking at astronomical circumstances somewhere other than where you are sitting.
If I ask Astrolabe about the sky over another city, sunrise at that location should be expressed as their sunrise.
Not mine.
There is also an option to show UTC when that is more useful.
Small distinction.
Important distinction.
Astronomy has an impressive ability to punish casual treatment of time.
I have accumulated evidence.
The data sources are visible too
One thing the recent work has made me increasingly particular about is provenance.
I do not want a page covered with impressive-looking decimal places without any indication of where the underlying information came from.
So the web version identifies the major foundations it is using:
JPL DE440 for planetary and lunar ephemerides.
IAU 2006/2000A for precession and nutation.
IERS for Earth orientation and Delta T where appropriate.
SIMBAD/CDS for stellar data.
IMO data for meteor showers.
CelesTrak for satellite orbital elements.
That does not make every observation infinitely precise.
Quite the opposite.
It makes it easier to be clear about which parts are calculated, which depend on observational data, and where the remaining uncertainty actually lives.
That is much more useful than simply printing more digits.
It is not an almanac lookup
This is probably the distinction I like most about the web version.
Every astronomical figure on the page is calculated by Astrolabe itself.
It is not downloading a table of tonight's Moon position.
It is not scraping an astronomy site.
It is not asking another service when civil twilight ends.
Those answers come from the engine.
External data is used where external observations are genuinely required or useful, such as current Earth orientation, satellite elements, weather information, and similar time-sensitive inputs.
But the astronomy itself is being done here.
That was an important design goal for the HomeSeer plugin.
It carried naturally into the website.
Why make it public?
Partly because I realized there was no good reason not to.
The HomeSeer version serves a very particular purpose.
It allows automation to ask questions such as:
Has astronomical twilight ended?
Is the Moon above the horizon?
Is a satellite pass coming?
Is there an eclipse or conjunction worth knowing about?
But plenty of people who might find the answers interesting have no reason to run HomeSeer.
The astronomy engine does not care.
So the web version gives the project another front door.
Someone interested in astronomy can simply use it.
Someone curious about the HomeSeer plugin can see what is underneath it.
And someone wondering why I have spent this much time worrying about polar motion can now see the consequences rendered on a webpage.
I am not certain that last group is very large.
But they are well served.
It also makes a nice test of the architecture
There is something else I like about this.
Astrolabe did not begin as a web application.
It began as a HomeSeer plugin.
Yet the astronomical engine has been kept sufficiently independent that putting another interface in front of it did not require reinventing the astronomy.
That is a useful architectural test.
The HomeSeer plugin asks the engine questions for automation.
The website asks the same engine questions for a human reader.
The astronomy does not need to know which one asked.
That is exactly the separation I wanted.
The interface is replaceable.
The astronomical model is not tied to it.
And yes, this project continues to escalate
The original thought was:
HomeSeer already knows sunrise and sunset. I would like a little more astronomy.
That became Astrolabe.
Then Astrolabe acquired:
Planetary ephemerides.
Lunar calculations.
Eclipses.
Occultations.
Satellites.
Meteor showers.
Atmospheric refraction.
Modern reference frames.
Earth-orientation data.
Relativistic corrections.
Weather-aware observing information.
More than five hundred automated checks.
And an ongoing disagreement with mountains on the Moon.
Now it has a public website.
At this point I have stopped trying to predict where the project ends.
That seems healthier.
Try it
The web version is available here:
https://astrolabe.cyberlogix.net
Give it your location and have a look around.
No HomeSeer required.
No subscription required.
Just the sky.
Computed.