About

Two numbers describe this site and they are not the same number. 223 448 objects are catalogued: each has a page, every designation we can resolve, and a row in the API. Beneath them 60 091 093 survey sources are stored — 36 909 335 stars from Gaia DR3 and 23 181 758 galaxies and quasars from GLADE+, held exactly as those surveys published them. They carry no page and no prose, because “Gaia DR3 66526127137440128 is a star” is not an article and tens of millions of those is not a catalogue. Each of them does have an address, because a row this database holds and cannot answer for is a row it is using rather than keeping.

Neither number is 1.81 billion, which is what Gaia DR3 actually holds. The three tiers below are the whole of it: what is stored, what gets a page, and what is offered to a search engine.

• Stored — every row of a catalogue we take at all

  When a survey is worth having, it comes in whole. All 23 181 758 rows of GLADE+ are here, quasars included, and so is every Gaia source brighter than magnitude 15. That magnitude is the faint end of a 300 mm telescope, the deepest instrument this site models anywhere, and it is the one place hardware decides rather than editorial judgement: 1.81 billion rows at what one costs here is some two hundred gigabytes before a single index, and this runs on one server.

• Given a page — an identity, and something to say

  A survey row becomes a catalogued object when some catalogue gave it a name or a number that resolves — Messier 31, NGC 5128, HD 48915, PGC 10435 — and when there is enough beside it to write a page from. 53 366 of GLADE+’s 23 181 758 rows clear that; the rest are a position, a brightness and a redshift, which can be looked up and cannot be read. A designation that is only the object’s own coordinates read back to it tells a reader nothing they did not already have.

• Offered for indexing — only what a reader could search for

  Every survey source has a URL and none of them is submitted to a search engine: each one carries noindex, so it can be linked, cited and resolved, and it will never turn up as a result. 60 091 093 pages of “an object at these coordinates” would be a doorway-page farm whether or not anyone meant it that way. Storing a row and publishing it are separate decisions, and this is where they part.

So the promise on the front of this site has a boundary, and it is this one: every object in the catalogues below, down to the depth this hardware can hold, and beneath it every row of those catalogues reachable by its own identifier. Where a cut had to be made it was made on brightness and on identity rather than on interest, and it is written here rather than left to be discovered.