[00:00.000 --> 00:12.640] Hi and welcome everyone. I am here today to speak to you about a little bit about Wikimedia's [00:12.640 --> 00:20.880] open source ecosystem. So I assume all of you know what Wikipedia is and maybe some [00:20.880 --> 00:28.080] of you know that it runs on a software that is called MediaWiki. So all the Wikis run [00:28.080 --> 00:33.560] on this software but there's also tens of thousands of websites around the world that [00:33.560 --> 00:41.880] use this. A very cool example is NASA is using it for some other projects. But this is sort [00:41.880 --> 00:46.640] of the core of Wikipedia and the other projects and it's of course something that is open [00:46.640 --> 00:53.520] source that anyone can contribute to but it's not what I'm going to be talking about today. [00:53.520 --> 00:59.840] Because surrounding Wikipedia in all the other projects there is a huge ecosystem of software [00:59.840 --> 01:09.680] tools. You can think of these as like third party integrations. People build bots that [01:09.680 --> 01:15.760] do edits on Wikipedia. That fight vandalism. There are machine learning algorithms. There [01:15.760 --> 01:22.860] are pipelines for data that then go to research purposes. There's a lot going on. So I'm going [01:22.860 --> 01:31.760] to be talking about this part. Just a quick word about me. I'm a software engineer with [01:31.760 --> 01:38.280] a technical engagement team. You can see we're part of the technology department and our [01:38.280 --> 01:45.000] team is kind of split into parts. We have the cloud services team, SREs and engineers [01:45.000 --> 01:50.520] that build services and platforms for all these tool developers. Then we have developer [01:50.520 --> 01:58.360] advocacy. They do a lot of things. They are writing documents and running outreach programs [01:58.360 --> 02:04.760] and doing everything so that our technical contributors can build cool stuff on top of [02:04.760 --> 02:15.480] our platforms and content. So just to give you an idea of the scale of this, we have [02:15.480 --> 02:22.880] over 300,000 editors that contribute to Wikimedia projects every month. Wikimedia comments, [02:22.880 --> 02:28.920] which is the project that is for free, media files, videos, images, so on. There are now [02:28.920 --> 02:36.920] over 90 million media files on there. And we have 1.7 billion unique devices that access [02:36.920 --> 02:44.760] Wikimedia against statistics for every month. So some of you may recognize at least some [02:44.760 --> 02:50.920] of these projects. Of course, Wikimedia is the flagship. There are other ones like Wikidata, [02:50.920 --> 02:59.480] comments that we just mentioned, Wiktionary, and many more. So yeah, we're going to take [02:59.480 --> 03:04.080] a look at these tools, ecosystems that I mentioned in the beginning, and we're going to start [03:04.080 --> 03:12.520] from Wikimedia. This is the thing that most people know about us. And from there, yeah, [03:12.520 --> 03:15.680] we're going to explore the tools that are community-created software that interacts [03:15.680 --> 03:21.840] with and contributes to Wikimedia projects in some way. An example here is PyWikibot. [03:21.840 --> 03:27.720] It's a framework for building bots. So if you have a wiki and you want to run some kind [03:27.720 --> 03:35.120] of bot that does something, some type of edits, you would very likely use PyWikibot as a framework [03:35.120 --> 03:44.240] to develop this. From the tools themselves, we're going to go and have a look at the services [03:44.240 --> 03:54.160] and the platforms that support them. And we're going to start with an example of a couple [03:54.160 --> 04:00.720] tools and how they actually integrate with one of the projects. So this is a wiki project [04:00.720 --> 04:07.360] called Women in Red. And a wiki project is a group of users on Wikipedia that decide [04:07.360 --> 04:13.280] that they want to work on something specific. They come together to work as a group. And [04:13.280 --> 04:20.000] in this case, it's to fight the content-general gap. So they observe that only around 15 percent [04:20.000 --> 04:29.600] of English Wikipedia biographies were about women. And as of the 23rd of January this year, [04:29.600 --> 04:38.520] they have managed to take this number up to around 19.45 percent in about seven or eight [04:38.520 --> 04:44.840] years. And so where does this very precise statistic come from? You can see mentioned [04:44.840 --> 04:52.080] here in red, it's something called human-niki. And so this is what we would call a tool. [04:52.080 --> 04:57.600] And in this case, this is a dashboard. It provides statistics about the gender gap on [04:57.600 --> 05:03.680] all the Wikipedia projects. And you can see here that female content is the orange part [05:03.680 --> 05:11.000] and then the rest is male. And if you go to this website, you can see it in a more granular [05:11.000 --> 05:25.240] way by country, by project, by date of birth, and so on. [05:25.240 --> 05:31.320] So if you want to contribute to this project, an easy way to do it is to go to this Wikipedia [05:31.320 --> 05:36.120] project site and you can see different lists that have been curated. For instance, here [05:36.120 --> 05:43.840] we can see female activists. So you can get a list of all the female activists. And there [05:43.840 --> 05:48.360] are many, many categories like this. And some of these lists are curated by humans, but [05:48.360 --> 05:56.960] most of these lists actually come from a bot that's called ListeriaBot, which curates, [05:56.960 --> 06:01.800] which makes queries on Wikidata, which is another one of the projects. You can think [06:01.800 --> 06:08.880] of it like a huge knowledge graph that you can query using a similar language to SQL. [06:08.880 --> 06:17.920] It's called Sparkle. So yeah, you can use Wikidata to get lists with very high granularity. [06:17.920 --> 06:22.840] You can have activists from Germany or activists from Germany that were born after a certain [06:22.840 --> 06:30.760] date. And so this is what ListeriaBot does. [06:30.760 --> 06:36.280] So we have seen two different tools. One was a dashboard, one was a bot. And there are [06:36.280 --> 06:41.880] thousands of these tools. There are thousands of maintainers. And we're going to take a [06:41.880 --> 06:50.520] look at how we sustain these people. So I mentioned that my team is the cloud services [06:50.520 --> 06:58.600] team. And what we do is we provide hosting. We provide compute virtual machines. We provide [06:58.600 --> 07:04.960] data services for all these tools to function. And so again, to give you an idea of the scale [07:04.960 --> 07:11.480] of this, 30% of all the edits on Wikipedia as of 2020 were made by bots hosted on our [07:11.480 --> 07:19.600] services. For Wikidata, that number is a little bit higher. It's around 50%. So just [07:19.600 --> 07:27.680] to make you aware that this is a quite important part of the ecosystem. [07:27.680 --> 07:34.120] So I mentioned there are thousands of tools. And as of a couple of years ago, we now have [07:34.120 --> 07:38.960] a catalog where you can browse and search and find the tool you need for your project [07:38.960 --> 07:45.400] or if you are a tool maintainer, you can add it here so that people know it exists. Then [07:45.400 --> 07:50.640] what you see here are lists that have been curated. We have something called the Coolest [07:50.640 --> 07:56.800] Tool Award. So if you look down, you can see that Humaniki was one of the award-winning [07:56.800 --> 08:05.240] tools in 2021. Some of you may recognize this as a Jupyter [08:05.240 --> 08:12.240] Notebook. So this is a Jupyter Hub deployment that we have that is directly integrated with [08:12.240 --> 08:18.160] all of our data services so that people can access dumps, they can access Wikireplicas, [08:18.160 --> 08:25.000] they can access a lot of things that otherwise would be gigabyte and gigabyte and gigabyte [08:25.000 --> 08:32.920] of data they would have to download onto their own computers. [08:32.920 --> 08:40.040] Another tool is called Query. It's a public query interface for Wikireplicas. Wikireplicas, [08:40.040 --> 08:46.280] I didn't mention it, there are replicas of our production databases. And the cool thing [08:46.280 --> 08:53.000] about this is that all the queries are public so people can actually search and see other [08:53.000 --> 08:58.640] people's queries and be inspired or if you're not very good with SQL, you can adapt someone's [08:58.640 --> 09:15.120] query to your needs. Here you can see a specific query. [09:15.120 --> 09:22.160] So these services are still tools that serve this ecosystem but we also need somewhere [09:22.160 --> 09:30.480] to host them. So though we have a platform as a services offering, it's called Toolforged. [09:30.480 --> 09:35.520] It's not quite as fancy as Heroku or DigitalOcean or anything of this sort. If you look closely [09:35.520 --> 09:41.640] you see that you have to actually SSH into it. But it's still very powerful and very [09:41.640 --> 09:49.160] convenient for our users. It integrates again with data sources and it has managed databases, [09:49.160 --> 09:58.000] a elastic search cluster that everyone can use without having to maintain all these systems [09:58.000 --> 10:05.120] themselves. So yeah, the back end here is Kubernetes. [10:05.120 --> 10:13.480] Then for more complicated projects, some projects need more compute for instance. We also have [10:13.480 --> 10:21.080] a CloudVPS offering so that people can spin up their own virtual machines and basically [10:21.080 --> 10:32.000] do what they want on them. So this runs on top of OpenStack. [10:32.000 --> 10:41.240] And how could one get involved with this? So it's possible to get involved in any of [10:41.240 --> 10:49.840] these layers, either as a tool maintainer or as maintainer of any of these platforms. [10:49.840 --> 10:55.360] And that's kind of the thing I wanted to highlight a little bit today is that this is kind of [10:55.360 --> 11:07.800] a unique opportunity to actually contribute to platform and to infrastructure. We have [11:07.800 --> 11:16.520] people that work with our team and they are on our IRC channels and they push patches [11:16.520 --> 11:20.920] just like everyone else and if you don't know you would think they are just another software [11:20.920 --> 11:29.720] engineer on the team. And I asked some of them why, what brings you here? And of course [11:29.720 --> 11:34.680] many of them associate with their free knowledge and free knowledge movement and open source [11:34.680 --> 11:40.760] and all of that. But many also said that this is a unique opportunity to actually play with [11:40.760 --> 11:47.360] things like OpenStack or Terraform or Kubernetes in a situation where actually you have real [11:47.360 --> 11:52.840] traffic and real users which is something that is kind of very difficult to do at home [11:52.840 --> 12:01.040] and there are not many other projects where you would have this possibility. [12:01.040 --> 12:12.600] So some ways to get involved. We have several outreach programs. We have Outreachy which [12:12.600 --> 12:23.000] is an internship that runs twice a year. It's targeted more towards underrepresented demographic. [12:23.000 --> 12:31.440] Google summer code that's once a year and both programs are open to anyone so you don't [12:31.440 --> 12:35.640] have to be a student. You could be someone who is changing careers or doing some kind [12:35.640 --> 12:42.080] of a letter move. Google summer code has also become more flexible. It's not just summer [12:42.080 --> 12:46.840] anymore. There are shorter projects. There are longer projects. So that could be a way [12:46.840 --> 12:56.640] to get involved and get some kind of hands-on mentorship. Another way would be to come [12:56.640 --> 13:03.320] to the Wikimedia hackathons. We have one in Athens in May this year and then one is [13:03.320 --> 13:11.240] part of Wikimedia that takes part in Singapore that is in Singapore in August. And of course [13:11.240 --> 13:16.880] if you are brave you can just dive right in because everything we do is open and it's [13:16.880 --> 13:25.880] out there on the internet. Documentation of course but just even our project boards and [13:25.880 --> 13:33.000] fabricator it's a collaborative software for task management and such. So if you go there [13:33.000 --> 13:38.480] you would see that there is a huge variety of tasks. You can see the work boards of different [13:38.480 --> 13:45.560] teams at the foundation. You can see volunteer led projects and projects where people work [13:45.560 --> 13:52.840] together alongside each other. So a way could be simply to find something that interests [13:52.840 --> 14:03.280] you and look at the documentation and then come on our IRC channels and contact us and [14:03.280 --> 14:13.960] that's it. So yeah I have added some links which can be helpful to get started. And you [14:13.960 --> 14:19.160] are of course free to just reach out to me. I had my Twitter handle on the first slide [14:19.160 --> 14:36.600] and my slides are published on the website. We have 45 seconds for questions. Thank you. [14:36.600 --> 14:54.000] Thank you.