The water side of the site grows up
If you have ever caught a ship in Skydex, you have probably ended up on a shipyard page. What you have probably not done is find one on purpose, because until today nothing on the site linked to them. Eighty-four builders had a page each, reachable from a vessel card and from the sitemap, and from nowhere else.
The aircraft side of the collection fixed this a while back. Types, airlines and manufacturers each got an index you can browse, and the difference was immediate: a page nobody links to is a page nobody reads. Water has been waiting its turn since.
Three new indexes
/shipyards lists every builder whose ships players have caught, biggest fleet first, with the country and founding year where Wikidata knows them. /flags lists the registries the collection has seen, with how rare each one is. /ports lists the harbours the watched fleet has called at over the last 30 days. All three sit in the footer of every page on the site, so the whole water tree is now one click from anywhere.
The pages say more
A shipyard page used to be a name, a link to Wikidata and a wall of ships. It now opens with a sentence built from what we actually know about the yard (where it is, when it was founded, what it is) and carries a photo, the country as a link to that flag's own page, and the list of pilots who caught its ships, first catch first.
Flag pages got the same treatment. Both now say who was first, who has caught the most, and when.
None of that text is written by hand, and none of it is invented. Facts we do not hold simply do not appear. Yards founded under states that no longer exist are a good example: Wikidata will tell you a 1905 yard belonged to the German Empire, which is true and useless, because that yard might stand in Germany today or in Poland. Where the answer is genuinely unknown, the page says nothing instead of guessing.
A page that can only exist here
The new block on a vessel page is the one we are most pleased with. It says which receivers in the Skydex network have had that hull in range, on how many days, and roughly where those antennas are.
Every ship database on the internet can tell you a vessel's length and her flag. None of them can tell you that your own network heard her on nine separate days from a rooftop near Hamburg. As of today, 161 of the 693 hulls in the collection have been heard by one of our own AIS receivers, and that share only goes up as volunteers add antennas.
Two things are deliberately missing from that block. It names the antenna's nearest town, never the ship's position. The daily rows we keep hold ship identifiers with no coordinates in them, and a station's own page publishes no finer step either. And it counts whole UTC days, not voyages: a receiver cannot see where one voyage ends and the next begins, so calling them voyages would be a number we cannot stand behind.
More pages, fewer hidden ones
The last change is the least visible and probably the most useful. Until today a shipyard page was only allowed into search results once two of its ships had been caught. That rule sounds careful and is actually a trap: the page is found through search, and search never sees it until somebody has already caught a second hull from that yard. Fifty-nine of our eighty-four yards were sitting in that hole, each with a name, a country and a founding year that nobody could find.
A yard now earns its place on what it can tell you, not on how lucky we have been. Two ships caught still does it, and so does one ship plus real reference facts about the builder. About seventy-five yards clear that bar today, against twenty-five yesterday.
What we looked at and decided against
Aircraft have a middle floor: the airframe belongs to a type, and the type is a page, a collectible card and the thing rarity is measured against. The obvious question is what plays that part for ships, and the obvious answer is the vessel class: the Triple-E, the Panamax series, the named classes people talk about.
We measured it before building it. Of the 253 hulls in the collection with an IMO number, 203 exist on Wikidata, and eight of them record a vessel class. Seven classes, six of which would be a page about a single ship. The broader "what kind of ship is this" field is there for all 203, but a third of them say only "ship", and nineteen values appear exactly once.
So there is no third floor in the data, and we are not going to fake one. The yard and the flag are the middle of the water graph, they already existed, and what they were missing was everything above.