 C'est-à-dire les vues, et si vous voulez savoir la structure d'une table spécifique, vous pouvez utiliser slash d t, nice slash d, ok, on verra dans les slides. Slash d, backslash d, name of the table and then you get the list of fields in one of the tables, so that's the package table, so you see that we import basically everything that is in the package files. Is it big enough? Probably. So who is trying that? Can you bring it to your hand? It's a low success rate. So then let's try a simple query. What's Ryan email address? See Ryan something, no? Something like that. I think it's really... It doesn't make anything or it's okay. So that's the list of fields because I use select star and then you have the data for each row, which is not really easy to read like that, so we can just do select, for example source version, distribution component, and then you get, well, let's use release. So you get each package in each Debian release, so it's maintained. Yeah, so as you can see we also import data for Volatile, for example. So then another example, which is a bit more complex, so let's look at orphan packages, which is nice because it's not a two big tables, so I'm quite sure that we want DOS, you did using that. So that's the list of orphan packages. So let's just reduce it, so I will remove the description. So there are columns for the source package name, the type of orphanage, whether it's RFA, ATI, O, and the bug number. I know what we could try to do is join the orphan package table with Popcon to get the most popular orphan packages. So, well, that's the number, well, which is the number of orphan packages. So there's a table called Popcon SRC, which provides Popcon for source packages. So Popcon is usually for binary packages. So if you want to use it for source packages, you have to, it's the most logical way to do that, to take the maximum Popcon value over all the binary package of each source package. So that's the value that we have there. And there you have the name of the source package. And so we can just do a query like that. So joining orphan packages with Popcon SRC, using O dot source equals P dot source, and ordering by number of installation. So that's a rather simple SQL query to sort that, well, it won't work anyway. So here we can see that there's a number of installations and the name of the orphan package. So as you can see with a few steps of SQL, you can really get information that would be really difficult to get otherwise. So that's the Evensburg that is still not fixed and stable. So when you add the slide before the last one and you type next, it goes to the end of the presentation and you have to go back, which is kind of. So as you can see, UDD already works. There's not so many things remaining to do on it directly. But we really need help now from people who would like to play with UDD, use it to build new tools using UDD. So we actually knew what is missing in the database or what should be improved. Because without using it, it's really difficult to discover whether you made a mistake in the schema by forgetting to import one kind of data or by not formatting it properly so that you can join the table. There's some work about importers that still needs to be done. So the first one is to import a wanna build, which should be a few hours of work at most because there are dumps provided by wanna build. There's also a Brittany, it would be nice to import so we know why a package didn't migrate to testing. And the MIA status of maintainers, so we know who is MIA. And the big thing that needs to be improved is examples and documentation. So I hope that I will find a victim today that will come to me and say, OK, I'd like to work on that. Because the point is, it would be really nice if someone quite new to UDD would do that. Because it's a really nice way to learn about UDD. So currently, we have some examples on UDD.dbn.org. But the general output is really ugly. So it's just text files. I need to go to the source to see the query. So it's just a matter of improving the rendering of all of this page. So to contact people about UDD, you can use ISE on dbnqa or the dbnqa mailing list. And that's URL where you find more information about UDD. Thank you. So any questions, comments, things that we forgot that we should import, stuff like that. Any victims that would like to improve? No, one remark, I really like this DEHS import, which is quite early new. But I just learned that the permissions are not correct. So I couldn't select from this table. Yeah, to select. Thank you. Anything else? So it depends on the data source. So, for example, for source season packages, it's a date twice a day for squeeze season experimental, once a day for older releases. For Ubuntu, it's also once a day. For various simple things like DHS, Popcon, Lintian, all it's once a day. Debian bugs, it's four times a day. Test immigration is twice a day, because we need rents twice a day as well, I think. And basically it's once a day for most, except when there's a good reason to update more questions. That one or the HTML page? The numbers don't match the Popcon numbers. They should. This one matches, no? Yeah, there's a small difference. Yeah, it's about the same, but it's slightly different. I don't know why. And maybe the, yeah, strange. I'll check. I checked that all the package, all the orphan package have a Popcon value. That's why I didn't use left join in the query, because, but, OK, maybe there's a bug. I don't remember how it works. Actually, no, that's not the one. So, it gets this file, I'll check, but maybe this file is no longer updated. Popcon was moved to another host, so maybe there's something broken in the process. OK, I'll check. Other questions? Can you think of something that we don't import yet and that we definitely should import? Oh, I'm impressed. OK, thank you.