 So this is the Media Wiki Stakeholders Group meeting here. As it is an open meeting, I think we do more like a QA, talk about what we're doing and what's all of this about. So there is no set agenda for this. What I did was I prepared three slides or four just to give you a brief overview of what Media Wiki Stakeholders Group is and what we do. And I see some faces I haven't seen before. So I think if you have any questions about the Stakeholders then this is the right place to ask them. And of course, we want you to get involved with us. So the question or basically what is this user group about? It is not responding to me. So that's taken from the website. So we are a user group, which means we are some kind of formal entity within the Media Wiki Media Movement. And we target, or most of our members are developers, system administrators, consultants, hosting providers, anyone who's involved with working Media Wiki outside of the foundation. So for third parties, which is the term we want to somehow avoid because not sure what is the second party or the fourth party or the first party. So it's basically we're targeting anyone who uses Media Wiki outside of the foundation's context on no Wikipedia and no Wiki source and stuff. Having said that, of course, we don't exclude people who also work on these projects. The opposite, we very welcome all of them. And so our goal is to improve the situation for these people. I helped them work Media Wiki, do a bit of lobbying also that the relevance of Media Wiki as a software product is recognized in the media world. We work on like campaign clients, rewrite extensions. And yeah, so there is quite a bit of scope we have. In short, we want to be a point of communication and coordination among Media Wiki users, meaning people who run Media Wiki. So that's a bit of a confusion sometimes. We are not specifically targeting anyone who is in touch with Media Wiki somehow. It's more about the issues of people running these systems. We want to coordinate different groups of Media Wiki developers and operators. And I would say more about that later. So also we want to foster and create an ecosystem around Media Wiki so that extension developers have an easy to contribute their extensions. If we try to work on stuff that targets especially these people, so that is not related to Wikimedia so much. So for example, we work on a thing called Adopt Integration or Adopt Authentication, which is not the main mission of Wikimedia, but a lot of Media Wiki instances out there use it to connect the user base of their organization to the user base of Media Wiki. Facilitate implementation of Media Wiki features, improve documentation, and contribute to the development of Media Wiki. All of this is what we intend to do, what we do in parts. The group is still growing, so this is like our mission. Maybe to show you a little bit of what we did in the past. So a few years ago, we conducted a Media Wiki user survey asking people how they use Media Wiki, what they expect Media Wiki, what the pain points are. Mark, thankfully, took over the maintenance of the API. So that has moved to the cloud services now. It won't be soon. It will be moved to cloud services soon. And yeah, so since the original maintainers, if I'm informed correctly, did not want to maintain it anymore, so we keep the service open. We do tracks and conferences. We do also conferences. So for example, we had this fantastic Media Wiki track in the Vienna Hackathon, but also the Enterprise Media Wiki conference in Houston this year. And I guess one of the things that's also a achievement of the group is that we have a good network right now. So the network is growing. And we have members of the user group that are in various positions of our stakeholders. So there are members of the foundation who also participate in MWStake. We have contributors from large enterprises. We have small businesses. We have consultants. We have volunteers, of course. And that's so obvious. I didn't put it on there, but I should for documentation purposes. So it's not only targeting professional users, as in people get money for using this. But of course, also people that use Media Wiki for their action groups, for their pet projects, for NGOs, for any kind of volunteer-based stuff. Currently, I think these are the most prominent projects. We see that this is not a full list. And I hope I did not miss anything. So one is, as I said, we are trying to get people together to do a standardized way of dealing with adult authentication. Probably that's going to be extended also to the summer at some point, because that's the new hype of authentication. Because we feel that this is a need, and we've seen a lot of separate efforts. And one of the things we want to do as a group is to combine efforts. So for two reasons. First of all, to save effort, of course, because there is like 10 entities trying to solve the same problem. That's just a huge waste of time, ideally. And second is to find a unified way to do this and find a good way to do this. That is sustainable. And we work with future MediaViki versions as well. Another project is MESA, which is a MediaViki deployment tool. And alongside that project, there comes a unified MediaViki installer. So we're working on improving the installer of MediaViki, especially in the light of the new setups, where you have a lot of services. You have to set up when you want to run MediaViki. So trying to make it easier. And I put one here, because I think that perfectly fits into the mission of what we do. That's the new privacy law in the European Union. And we're currently trying to find a position of how MediaViki relates to that GDPR. And that would probably help a lot of people out there using MediaViki in their organizations. So how can you get involved? I guess the main entry point, if you want to get in touch, get involved in what we do is go to mediaviki.org and MediaViki stakeholders group. You don't have to remember all of this. There is a simple solution you can search for MWStake, which is our short abbreviation. And you will be redirected to that page. So MWStake is what you want to keep in mind. And here you see a lot of entry points. You see a list of current projects, not updated for the GDPR there. You see the communication channels. And that's probably one of the things that are a bit obscure about our group. So we don't go to the mailing list so much for, I think, historic reasons. It just so happened. And we try new methods of communications. We had a select channel before. Now we are on Riot. And a lot of the project work is done on the Riot channels here. That's what you might be interested in. There's a high frequency of contribution. So we have several conversations a day on these channels. And that's also the point where people get connected. So on these channels, we have the multiple stakeholders I talked about. So there are people from the foundation. There are people from large enterprises. There are volunteers, and they all work together and try to solve immediate problems or issues. On this page, you can also see, as we are a formal group, we have to do some kind of reporting. It's quite easy. I have to do, I do this. It's not hard because we don't have any finances involved yet, so we just can compile a list of things we did. That might change at some point. Yeah. So apart from these lists on Riot, there's also a monthly meeting we conduct, which is done. It's a video chat session in a system called Blue Jeans. That's not hard to set up. You just follow a link, and then it opens up. Because Chris always sets up the sessions for us. And that is a task on fabricator. I opened it. Now, where is it? How do I do this? So we have, how do I get to the calendar? I can have my video right there. We'll go ahead and do a list right there. Let's do this. So we also have a website, mwstate.org. And there we have the meeting. There is no fabricator here. Yeah. I'm going to zoom to it. It's going to from here. Really? Yeah, well, I'm sorry. You want the. The point is it goes right to the Blue Jeans link at the bottom. Yeah. Click that link. OK. OK. I clicked it. So I'm usually invited by Mark personally. So I don't have to know what the solution is. Yay. OK. That's a monthly task. And you can see the previous and next buttons here. So you can always find the next meeting for this. Yeah. So that is the end of my fabulous presentation of the stakeholders, because I think most of you know what we're doing. And now for me, it's like open. So the question is, for people who are not actively involved here, what would you expect of the group? Do you want to participate? Do you have any pressing issues about media that you want to discuss, which we should tackle? That's also a question. So for example, one thing we are heavily discussing right now is, again, extension management composer. That's also one of the things we try to facilitate in some ways. And there is a session at 7 o'clock tonight. I'm not sure which room it is. But to discuss the extension management or I think they can all post it. Exactly. With sign bar about composer, right? With the sign bar about composer. Well, you've been talking about composer. Yes, I have. We can talk about composer. I think it's right in this room. Is it here? Let's see if I can. Here it is. Here it is. So you only have to stay until 7 o'clock. It's GDBR. Yeah. So this is basically, if you stay here, you are virtually practically involved in the stakeholders. Yeah, so are there any things that, I mean, are you all users of Media Weekend and what bugs you most about the software? You said I was interested in the idea of integration. You said it has to be moved forward now. But I also know that for example, I thought of all about form was working on it a bit. What part of integration? From other parts. Yeah. It's the same project. It's the same project. Is there any, what's the schedule? When can we use it? We were just talking about that before this. You're talking about that, right? Yes, sir. Yeah. We were just talking about that. Just out there, Robert and I were working on it. Just now we're trying to merge some changes that I have with his work. I need to deploy it for customer of mine. So there'll be soon. OK, right. So one of the questions about it that I think needs some discussion and socialization is the naming. So there is an existing Neldaq authentication expansion that still sort of works, but it doesn't. My understanding is the issues that it has to do with the single slide on. I think that's no longer working ever since we even did one of the 27 Neldaqs taken. So some people are able to still use some of the single functionality. So one of the, so the idea behind this LDAP rewrite is, first of all, to get a fully working version of LDAP authentication again. But also to create a uniform interface to LDAP services that can be used for querying the directory and to get user attributes and groups, as well as what are the other extensions? So the rewrite of LDAP authentication is on top of my plugable raw extension, which also differentiates between authentication and authorization. So I already have an existing LDAP authorization extension that will work with the LDAP authentication. So this is a complete rewrite of LDAP authentication. So the existing LDAP authentication uses camelheads. And we were thinking of reusing the name, but with LDAP we can categorize. But I do not, but yeah, exactly. And that seems like something that means some community of code just to take over the names. We could take over the name LDAP authentication with the camelheads. All of the other LDAP extensions in this week use all that all of the Facebook. We could also switch them, but the curious and mean things that's wrong, we could just LDAP us all. That's a fact. I could be pragmatic. But yeah, I use the same thing that, you know, just having a name differentiated in the case is completely wrong. Yeah, I agree. No, I would say that doesn't make sense. But if you say the old one is discontinued because right now you wrote the old one and has the old one you maintain it, take it over, but fix the spinning the sex of it. That would be one. But I also want to make sure that existing users of the existing LDAP authentication are comfortable that the new one works for them. They're going to have to change, update their interface. It's going to be different. The configuration aspect of it. And reusing the name means the other one goes away. It's currently not maintained anyways. Much. There's no pact with me. So just as an aside though, the current one is deployed to the Wiccumedia Foundation with you. So it can't 100% go away. Unless the Wiccumedia Foundation chooses to use the new one. And not likely to have had anyone just around to doing that. So this is why I raised the issue because the naming does have all sorts of implications. So the current, the old implementation is basically a patch wrote written rather hastily after the authorization system changed. And I don't think it's really stable in the ways that you can see it. So it doesn't fully support the old extent of the LDAP authentication. It's at least to the standard. It's in Wiccumedia's uses. Yeah, yeah. I'm just saying can't delete the extension. Yeah, right. It's still easy. But the question, I mean, speaking rationally, I think eventually even the foundation should move to like the extension that's even work for us. Hunt, or move to WordPress. Yeah, rewrite in Java. Ideally I can never be using all that. But that's the different digression I guess I shouldn't go into. Yeah, so we can't necessarily deprecate the old one. It's probably not a good idea to differentiate the new one and the old one only by the case of the LDAP, you know. Yeah, because Windows. For many reasons, including which tools like, yeah, I don't know, we'll even. It's fine. It depends if you're using Windows or Mac or Linux. It's Windows. Windows caused a problem. I think Mac actually is case-incessant too. It's case-incessant. Yeah, so whatever case you request is what case you'll get. But what did they do to bring this in for a minute? I know. It's like brain damage. Yeah, so that is not a good idea. But having an LDAP authentication and an LDAP authentication too is not necessarily a good approach either. So. It is better though, because at least in the case which one comes after, it's only maybe one of them gets dropped. I know it gets picked up. And then why is there two? And yeah. I have a lot of negative experience with dynamic page lists one, two, and three. Yeah, I was just gonna say that's the other example. If there's something about it that would actually distinguish it in another way, like how it goes things differently or something like that. LDAP authentication. But does it run everything? There we go. I think we're gonna just solve the problem. You just wanna push your brain. You want my word. To be fair, that might actually be a good solution. Perhaps. The old one's not maintained as a reason like why not to take over meetings and have one and kind of change it. It's a different architecture. Well, there's no reason that we couldn't just write our stuff on top of it. Basically delete all the old files and get and add all the new ones again. But as Brian says, it's a different configuration. And so if the foundation continues to use the old version. Foundation will not be happy if we delete something that currently uses it. Yeah. Yeah. I mean, as I say, but also if no one's maintaining it, I'm sure that they would be happy that someone's now actively attending that extension and you could make it kind of backwards compatible these for a bit and then kind of push it over. You get to talk about making it backwards compatible and working on that with a transition script, if nothing else. That you run a script and it converts from that. Right. And then you don't even need one extension. So that's a possibility. You know, there's still that one here, but it doesn't like it back to those sessions. But I can really affect that one too. Let it go. Let it go. I was really good tech that we're trying to find maintainers and you're obviously invested in LDAP. So it seems like a natural fit. And so the suite of extensions would be the LDAP provider, which is the glue that all of them use. LDAP authentication, LDAP authorization, which integrate political law. LDAP user info, which gives you a query LDAP, get back properties and then LDAP groups, which allow you to assign media wiki groups based upon LDAP group information. Am I missing any? I don't think so, but one reason that I didn't want to continue maintaining the old one was because it is old and it has grown and evolved organically, that would just, there's a lot in there. So while I often advocate taking over other projects and maintaining them, this is one case where I'm like, this is too much, not everyone cares about groups, for example, and so why should they pull in the whole, all of that code? Yeah, so it's easier to document and set up if it's split up a little bit. Also extensibility, so if you do integration projects, more often than not, you come across things that are not foreseen by the extension. So I don't know, a user wants you to put the organizational unit they have in the LDAP information on the user profile page. And it's rather hard to extend the old extension because of its architecture. I'm not saying, I mean, it's a very good piece of software or it was at the same, and my name was one of the best developers we had around here. But since like five years, the architecture of PHP or the state of the architecture has evolved. It's no longer a state of the art. It's a lot of spaghetti code. And authentication infrastructure within the user. The infrastructure has changed and now it's way easier to move forward to move forward by rewriting instead of improving it. Yeah, I just, this is kind of like, this is one of these things that seems like a vicious cycle of like we let something get old and then we rebuild it and then we don't get people to migrate and so we just keep building up, building new versions of things. It's like, you know, eventually you have to like build it and not build it. So I think that, I think migrating and reusing the old name would be great. What I'd want to do though is go through a beta testing phase to make sure. And so the question is, does anybody here, you know, we certainly know that it's used within the foundation. Does anybody else use LDAP, for an LDAP authentication that would be willing and interested in testing for you? I can do it. Well, you and I already have done, you know, and I already have it, I have an instance that has to do with the foundation of that. Oh, that's Cindy, we will. For anybody who has a, would be tech account, can log in after the test for you. You can, like, develop as a branch if you really want to actually take over the extension like you first have it as a branch and eventually place it past your tech account. If whoever wants to continue using version one and maintain version one, can you? You could have us leave on the branch and we then work with the auto branch and we need to release it. You don't have to. Yeah, that's the right answer. You can leave that yourself. I couldn't leave the answer, that was before I asked it. But. Ah. Yeah. It kind of depends on how the foundation updates the extensions to the variables and branch spaces. Like, when we, we don't use the release branches, but we do make, every week we make a new branch based on the master. But I don't think that's written in stone in the past. There was like central to this recent section at one point. Like, I think there have been exceptions in the past where I've heard a lot of people be like, yeah, maybe it's going to be a branch or something like that for the master. It sounds like it's a problem that's coming up, but yeah, I guess I have more branches in the description. Yeah, I'm sure it's like a two line shell structure. Yeah, I think it's the thing that comes up. Yeah, the idea was like, if this piece is actually a new branch, is it basically a new branch? Yes, ma'am. So that would be relatively easier. So if we can test. Then what do you need to do, Germany would probably use this as well, actually. I don't know what you're thinking about it. So we know somebody who can tell us what that's that. But they're not a patient like you, we don't care no more. But it could be actually testing. I assume they use it. Yeah, so that's where I'd really like to go with the stats is to try and have it test just as widely as possible. Any other? Great, so which one is the old one? LDAP authentication, capital L. And what's LDAP authentication plug-in? Or is that the old one? Or is that something even? Yeah, I don't even know what that makes sense. Which is unfortunate, because if I want to call it a little LDAP authentication, if there's something already called LDAP authentication plug-in, that's sad. Delete it. Just read your KOS link, you can see if anybody complains. I think the new one should replace LDAP authentication. There's lots of words in English that offer that, which are really important. That's LDAP authentication flow. LDAP authentication may be changed. Yeah, I do think it gets confusing when you have too many extensions that do the same thing with slightly different names. It does. Like SimpleSanwell and SimpleSanwell. Or SimpleSanwell offices. Well, yeah, but then there's those people who name their extension after the software that they're using, like SimpleSanwell, PHP. What, how am I supposed to do that? Who would do that? What are we talking about? That's why I would name it SimpleSanwell. That's because SimpleSanwell off is already taken. Yeah. It should have been Plug-BullSanwell. Plug-BullSanwell, you see? SimpleSanwell. Or just Sanwell. Yeah, take it over. Okay, I didn't ask you guys when I needed it. All here's Sanwell and all of you else. Yeah, so that's the one thing is the LDAP. We've got a session on DAPRO, we've got a session on extension management. It would be great to get everybody to join us on Riot because we do chat all the time. Yeah, and have a chat out. Absolutely. Yes, but the audience is silent slack and we're getting rid of Slack. We're getting rid of Slack until EMLU is on a Houston. And yeah, this. And we're also serving as a test bed for Wikimedia because there's no space. Wikimedia will use IRC until it dies. Yes, and there will be a bridge to it, right? It's just that the bridge doesn't work really well right now. No, because if you have it dedicated to it. You have a fine RC channel, so we might be able to use it with you to both spell. I don't think the relay will come to pick things further. That's a point here. So the reason why we are not using IRC is because we want to connect with the outside world. And it's really, really hard to convince someone from the shiny organizational world to use IRC or to use IRC client on their computers. And I guess what Mark said, and we can also serve as a test bed here, I think the migration from Slack to Riot was rather seamlessly quite well. Migration, that was just user migrating, not conversations or anything. Yeah, yeah, yeah, but I mean the switching over for users to the other systems. Yeah, and Riot is not quite as shiny as Slack, but it's more palatable to us. So if non-Chinese users are trying to use it. Here it goes. It's working on, he has a ticket to spin up a Riot server for the foundation so that they can get better throughput tires. Is it really a throughput issue? Like these protocols are actually literally, they literally have one machine for pre-node. Protocol made in like 89 has throughput issues. No, no, it's not, it's not the protocol, unless Riot was. So it's not IRC's fault? No, it's not IRC's fault. Yeah, not at all. It does weird things like, when a new person joins an IRC conversation, it takes a while for their account to be recognized by Riot, and so their messages don't wind up being bridged, and so all you see is one side of the conversation and you're wondering why this person is talking to themself. So, which is really not good. So we have one room that bridges the media wiki L. No, obviously, the media wiki. In general, I actually try to open it so the room doesn't exist. Oh, really? Yeah, I followed the link from the stakeholders page. Well, it was very well-readed to the stakeholders. Media wiki, generalMachis.org, doesn't exist. Oh, I'm in it. I just joined it. Oh, so maybe it's private. No. Media wiki, let's see, let me look at it. It's something I don't touch more. Media wiki, generalColonMachis.org. Oh, no, wait, that's MWStake, general. Yeah, it's media wiki. It's just town media wiki at colon matrix. Yeah, but the MWStake general is a different channel. Right, it is. Yeah, MWStake general. Yeah, that's working. Yeah, the only one that bridges IRC is the media wiki. Just media wiki. Media wiki channel. So maybe it could be that the link here is broken. Yeah, let's see. Yeah, it seems it was broken. So the general part is the media wiki. Yeah, the general part is the general tech channel. But that one's flaky anyways, so I'm not necessarily. It's a wiki editor. I don't have my left about, but Marcus is going to do it. So that's what the general MWStake looks like. I didn't read, so if there's something that's making you laugh, I don't know. I don't know. Pardon? It's the interface. That's because I'm a native German speaker. Oh, really? Somehow. Not really. I speak a dialect, so people, native Germans, don't understand me. So yeah. So yeah, there's a bunch of rooms which you can see listed on the page down there, which I'm beside by. Right, and we'll take off that one after the session. Yeah, we're just going to move the general. People want to try. We might want to put a banner there saying it's a little flaky. But the hope is that once we spin up a server. Yeah. So then you can see that we do have some participants there. The thing about the riot bridge and all that is we are talking with one of the product managers, I guess, at Matrix, and he's very responsive. Yeah, there's a channel that we created there just for communicating. So some of the Matrix developers seem to be, they like this page. Hey, Wikimedia is a big organization. They're using it. Yeah, so I get help. I'm sorry, now you have to pick me up. What's the relation, the connection between Matrix and Matrix? Matrix is a protocol. Riot's the last interface. Instead, XMPP, which was XML, message is back and forth, Matrix is JSON. So it's cool. So it's cool. It's JSON. It must be good. But the other thing that would be great is to have folks dialing into the monthly meetings. Do you want to talk anything about incorporation? Yeah, well, I was maybe a few sentences. I was trying to keep this out as a major thing. But of course, we do have internal stuff in the music group as well. And when it's becoming a formal body, those of you who follow the group know that we have this plan for a few years. And still, it's on our to-do list because without being a formal incorporated, like, leanly entity, we cannot process money properly. So currently, the end-up mistake, as I said, is somewhat money-less or money-runs to private personal. So one idea that, and we're working with Brian, and Brian is a guy in California, and his wife is a CPX, so she can help out with this incorporation. One thing that was suggested was that we designate, because we had this problem with EMW Khan in Houston, that we designate someone to hold an account. If it takes us any longer to get incorporated, then we designate someone. You're going to manage the account. And so join in. WSA cannot be able to hold our money. I'll probably invite it to hold me. I want to hold your money. Yeah, but the deal is that if the organization sponsors a conference like EMW Khan, and there's any funds left over, we would like to be able to park them somewhere in the interim until we have an opportunity. They could benefit the next conference. We could, if there's enough money, have a small hackathon. The work that's going on on MESA, we keep talking about organizing a MESA workshop. So that is one thing, the grant thing. We have been talking about getting grants from the market. Right, Marcus? So I recently could apply for a grant. There's the grant that was the NASA gave that I'm going to be working on, and hopefully WMF will also contribute. So that may also lend us legitimacy. You want to talk about that project as an immersed in project? It's mad. Multilateral asynchronous bidirectional synchronization. It's a media Wiki remote forget. And there's people who are interested in getting media Wiki, a working installation of media Wiki on the space station, for example, that could sync up with Houston. And then there are also people who like WikiFindy, takes Wiki, the Wiki experience, the Wiki editing experience to disconnected places and tries to get them to contribute back to Wikipedia. And right now, they have all that. That's fine, but there's no automated way to sync up. So it's cut and paste and a lot of manual work. So you lose attribution of edits that way. So this would solve that. That's it. Yeah, so NASA's interested in forgetting the copy of both Wikipedia and they're also internally used wikis up on the International Space Station. The first step in that is just to get read-only copies up. But the goal is to, the next step would be allowing updates to be shipped when things change on the ground version. But ideally, you'd also like to have an astronaut who comes in from a space walk be able to update the wiki with information. Because that's what they use their EBA wikis for, is collecting information about how tools are used. So the best example that I can, the IO of how you would update the wiki from a space walk is the Sharks, where they have a catalog of sharp points around the space station. That is not often updated. But if you have a wiki there that you're updating it, the guy can just go in from a space walk and he says, oh, I found a new sharp. And that's stuff that could be critical. Yeah. So the idea to use, to your ideas that you used to get? For the like, basically? To do the synchronization, yes. That's already working. Oh, it's already working. But there's a lot of manual stuff that has to go on. So I'm working on proof of concept that allows you to do it without all the manual stuff. But it is just a proof of concept. So a couple weeks ago, Darren Welch got funding from NASA to fund their work on this space station application. And then. But just a proof of concept for it. I want to make everything, you know. And then Mark's looking to fund his contribution to it. But I think it's fascinating, the wiki funding aspect, that the same technology that could potentially help to synchronize wikis with the international space station can also be used for folks who want to be able to contribute from remote villages. That's awesome to me that the same solution can help those two very diverse use cases. So is it a get repo that the entire MySQL database is in? Or do you have an individual get rebos for each article? Or how are you doing? No, as a matter of fact, it's interesting you mentioned that. SMWCon in Frankfurt, there was someone who had set up individual repos for every page. No, it's not that. The remote is basically, the API client goes and gets the changes from the media wiki API, stores them as revisions in Git. You could actually check it out onto your own desktop and edit the violence in a text editor and then check them back in and then push them to the wiki that way. OK, so Git's just a transport mechanism rather than like. Right. OK. Transporting, it helps with conflict residences. I can't say any conflict residences. OK, yeah. How are you going to conflict resolutions? Are they going to be handled in the media wiki UI? That is ultimately the plan. That is not, you know. So probably from my pilot up, I'm going to say, you know, just accept everything from this side and they win all the time. OK. But yeah, eventually you want to have a UI where you say, OK, well, one thing Darren suggested was we can use templates some more. Are you familiar with Git? How Git's are. So instead of the Git-Git conflict thing, you would say, this put a template around the top part. This is from that one template around the bottom part. This is from us. And then, you know, you could look at them manually and know which one was which. That would be the right category or something, saying that these are the pages that are not conflict resolved. Yeah, I know, it's great. I mean, I often wonder too, I just like this is like, please guys step in. I often wonder about like, since Git is like such a good paradigm, like the developers use all the time anyways, and we are doing like revisions internally, I just wonder if there's like, it's worthwhile considering like, should we change in the entire way that we're storing articles to be more like Git that way, like something like this, you just get for free, kind of, because like, if everything worked internally with Git, like Git just automatically has all these capabilities. You don't have to talk to someone besides me about that. Yeah, I know. There was a discussion on the Logitech emailing was a long time ago about that. And I believe there was some concerns like naively doing that would not be performing like the performant path for Git and perform assumption to Git are very different from performance assumptions around how Wikipedia is used. And especially for, if the idea is extended where it commonly is to have like, everyone working on their own thing and then merging it all together in like a massive distributed manner. I think it tends to get slow when you start having very big results. Yeah, they do. Yeah, it's like, I mean, there are ways you get around that like you have to do like, be packing it and leveling like the... I'm sure there's a lot of smart things you can do but like that naive approach is, from my understanding, probably not the core goal. But you could probably fix that, right? So you sound like, you know a lot about Git, I might have to ask you some questions while I'm... Well, yeah, I mean, yeah, I played around, like most of my development would always use Git. Like I used Subversion and I hated it. Like eventually everybody started using Git which made everything a lot nicer. Well, that'd be subversion. What? Yeah, just like checking out everything and locking and no one else would be able to do anything, which was like crazy. Oh, yeah. Yeah, the locking, that was a stupid thing. Although I do remember working on something, having to do something else, so making a giant patch file for myself just so I can reapply it like 10 seconds later. Ah, true. But it was simple. Actually, what I missed most about Subversion was that before we had the code review, it was just so easy just to put my sheet code in and then it was all good. Well, yeah, but then I did that a few times and got chewed out. Maybe you just, you're kidding me, I was as bad as mine. So should I take this as the final note of the session? We should opt for, we should argue for migrating to Subversion again. Okay. I think we are, we should leave some minutes for the room to change. The only other thing I wanted to mention, briefly, is that we didn't talk too much about, and it can be a future discussion, but Meza, the installer that NASA's working on, the other NASA project. And there is a lot of promise in that, but also a lot of work yet to be done. So that's another project that we've come out of. So how can people contribute this way? There's a Meza room on Riot. Can you explain how it does it in 15 seconds? Ansible. It's a set of Ansible scripts. You've got two minutes, that's not funny. It's a set of Ansible scripts. It started with, and they run on Red Hat and Centos, and we have a project to do a Debian port. It incorporates, you know, they built their stuff. I gave them feedback based on the Ansible scripts I built. It's a picture farm aware, and so very easy to set up and make it, you know, it keeps this partition. You can have shared, you know, global extensions as well as per wiki-initial extensions. The limitation, I would say, at this point is that you still need, it's not for the faint of heart. You know, it's automated, but it's still not as simple as it needs to be to get folks to be able to just, you know, take a couple things and boom, stick that at a wiki farm. They haven't mastered the point and click yet. It's not even close to point and click. So the one thing I'll just tack on to the end of this is over the last six months or so for wiki days and wiki data, we now have a collection of dot images. That's who Lex needs to talk to. You need to talk to him. Yes, we have a collection of dot images. He's your doctor adventure. Perfect. For the query service, the query service of data, media wiki with wiki days and some other extensions which you wouldn't want for wiki days. And it can be as simple as if you have a doctor downloading one file, changing a couple of passwords so they're not the defaults and then just running the command that you have it all set up. And so for easy installation of wiki days, that's sort of the pathway down. Yeah. Is there any? So the thing I was worried about with the doctor and I know it's silly, is gender lock-in? Yes, we get that. Gender lock-in. Start, start, start. I don't know. I don't know. Let's do a couple here. So for example.