The official Ultimate Debian Database only allows Debian project members to access it. Mattia Rizzolo and I co-maintain a mirror. Query and enjoy!

psql "postgresql://udd-mirror:udd-mirror@udd-mirror.debian.net/udd"

I wrote up its history recently: https://announce.asheesh.org/2020/11/the-ultimate-debian-dat...

More info about the service: https://udd-mirror.debian.net/

That is cracking! Thank you. I note rather a lot of Ubuntu themed tables which will come in handy as well.

I use Arch for my personal gear, Debian for Home Assistant instances and Ubuntu for nearly everything else. A slack handful of Gentoo boxes hang around, looking moody and causing trouble 8) This sort of resource is a great idea.

I can see myself wiring it up to an automated change tracker or similar idea still boiling around in my head. Bugs are in there as well and a lot of other good stuff that should mean that I can run a query to show me what will break before I do updates!

Just curious, how do you configure the connection limit for this public postgres instance? Is it just a high value, or something more clever?

Ultimate package database, I think? Granted, including maintainers, bugs, metadata, etc., but centered around the packages.

I'd say a distro like Debian is nearly tautologically defined by it's packages.

I suppose that's true of most distros (the partial exception being distros that are also defined by their package managers, ex. NixOS and Fedora Silverblue). That said, the packages don't cover everything around a distro; off the top of my head, it only partially covers people (yes maintainers, but people could be involved in other ways), and it appears to not cover events, funding, or documentation (ex. wiki stuff). I'm not saying that's a fault, mind; even the existing database feels like a weird commingling of concerns to me, but <bikeshed> I wish they'd called it the Ultimate Debian Package Database or something </bikeshed>

What is this supposed to be used for?

I'm seeing it as simply being open about internal workings mainly. I assume it is an internal database used as part of managing the project and Debian has always generally been a project that abhors keeping things under wraps.

For someone wanting to be more deeply involved than simply having Debian installed, access to a copy of this database could be a valuable learning and/or reference resource.

More abstractly, for someone learning/revising relational database concepts it could be a useful structure & dataset to practice on, particularly around the package dependency graph (a lot of example datasets out there as learning resources are hierarchical or just plain flat, so this is a welcome addition to that arena), and if they are at all interested in Debian then it is an example dataset/structure that is more relevant to them than a lot of other examples would be.

The wiki page lists some of the use-cases.