Wikimedia Developer Summit 2016 opening keynote by Wes Moran: transcript

Key points

• Moran highlighted 2015 work including single-user login, HTTPS, the HHVM move, performance gains, a rewritten pageview definition and the revision evaluation service.
• He also cited RESTBase, VisualEditor expansion, and Content Translation producing over 40,000 translated articles.
• Top 2016 priorities include mobile editing and speed, API efficiency, cross-wiki features such as notifications and search, and data and feedback tools.
• He said the main ways to engage are team Phabricator boards, wiki pages, mailing lists, talk pages, IRC, goal pages on MediaWiki, product showcases and a product hub.
• He said planning is to move earlier, with drafts for each quarter prepared sooner and a goal of a 9 to 12 month roadmap.
• In Q&A, a central global repository for templates, modules and gadgets was called the most important community wish list item.

Chapters

1. 0:00Welcome and purpose of the talk
2. 0:48Highlights of 2015: performance, analytics, research
3. 4:54VisualEditor, translation, notifications and mobile
4. 7:21Priorities for 2016
5. 9:31How to engage with product teams
6. 11:46Planning cadence and roadmap
7. 13:01Community wish list and the community tech team
8. 16:26Q&A on prioritization and shared templates

Full transcript

0:00Good morning. Fortunately, that cut into the time that I'm supposed to have already, so we're going to make this really quick now. I'll be speaking for about 20 minutes. I left some time at the end for question and answer, hopefully, and mostly I just want to thank you all for coming. Like you, I'm here to have discussions. I'm here to participate and hopefully even build some stuff. What I wanted to try to do with my time today is really communicate a little bit about what we're doing at Wikimedia Foundation and what we're doing with the product and the technology teams, what we've done, what we're going to do next, and really how to engage with us.

0:48So there's a lot to celebrate, I think, about what we've done in the last year. There's a number of initiatives that we've done that involved a lot of people. There were multi-year efforts that involved a lot of people that pays benefits for everything that we're trying to build and do. The single-user logon was a many-year effort, just like the HTTPS was. This year, they came to fruition, and... They are both things that went pretty well coming off of the multiyear effort. The hip hop virtual machine is also something that we did this year, which greatly helped reduce the editing experience and the load time by 50%. We also decreased the paint times on page load times by reducing them by 40%. This was done by the performance team.

1:48performance is really important it's it's where we all benefit it's where users benefit the most and it's a an area that we want to see improve and expand this year as well the pageview definition that was rewritten this year was also very important the analytics team helped redefine making the pageview definition go from a sample blog into a unsampled log which gives us a much more accurate description of what's going on across our site.

2:20Turning that into an API has allowed many teams to start using that for development, which is a very good thing as well. The analytics team has done a number of dashboards with the different teams and enabled more awareness through extending the event logging. the Objective revision, I'll just say ORRES, the revision evaluation service is really the machine learning service that helps us get much better quality across our entire set of software.

2:52This is something I hope to see expand into a lot of the different features as we develop over the next year. And then the release team did a great job really getting everybody on Fabricator this year. as product teams that's the best way to engage with us and you'll see a little bit further we have that's the primary way to actually engage with us the research team has done a dramatic amount of work this year one thing that I'd like to call out is the new editor retention initiative that was done through tea house the reason why I bring that up is because I think it has a lot of potential, and it's highlighted here, but I think these are the types of things that we have to do to move product further by engaging with our users more directly.

3:50There was also a number of design research initiatives that were done this year that really, really, really help how we're developing our software, through A-B testing, through surveys, and that paid a lot of dividends to the visual editor work that went on this year. Research has also done a lot to open our data sets and provide a lot more information for people to gather and to understand what's going on. Restbase was rolled this year. There was a lot of improvements that came from that work.

4:20both in the save times, editor switching, and the reduction of our overall HTML load. API performance in general has been an initiative where we are looking much closer at the performance and the analysis of our APIs, and the simplification and the documentation around them. The improved reliability of Wikimedia Labs was an initiative done by the operations team. There were some issues earlier in the year that I think they recovered and have done a much better job of actually improving the lab service.

4:54Visual editor continues to expand. A lot of features were included this year, including the math extension, improving upon the quick loader, and the incorporation of the graph editing, I mean the graph extension. there's a lot to be A lot more is going to be done over this year, but we really have started to expand now that we're on English Wikipedia. The content translation tool resulted in over 40,000 new translated articles this year. This is another thing that is providing a lot of new opportunities for us to engage more users. Echo notification enhancements, I like them personally, I'm using them a lot.

5:49And we're gonna have the ability to do cross-wiki notifications very soon. Parsoid improvements around HTML size and performance were also done. Flow beta was released. The graph extension itself has gone through a lot of iterations and now there's dynamic graphs, which Yuri has been working on that show a lot of promise. Search completion suggestor is working on improving the relevancy of how people actually discover all the content across the site.

6:19The map tile service and the wiki data query service went into limited production and this was an opportunity for us to work with operations to really explore how to expand things out of labs in a more bare metal way. And there's probably too many highlights. There's a lot that I'm missing. But I want to try to make sure that we celebrate the things that we are doing and the progress that we are making. We've reduced page load time in mobile web. There's been a lot of Android and iOS additions this year. The tab browsing and the link previews as well as the map service showing up in Android are all great things. iOS with Share a Fact and improved search as well as a read more section all have opportunities to be showcased in the desktop as well and that team is continuing to explore those the banner history feature that was entered into central notice allows for better understanding and testing of how banners perform and then last but not

7:21least the community wish list survey that was introduced late in November has provided a better engagement opportunity for the community tech team to do directly interact with our community to find the primary things that we want to do to actually increase and improve our admin tools. I'll talk more about that later. So what are we doing next in 2016? These are just a few things. Being here and going through our annual planning, the constant feedback loop with individual teams will expand and improve upon everything that we're trying to do over the next year.

7:59we're Mobile editing and mobile speed improvements are at the top of our list. These are things that we absolutely must do in order to try to improve our overall software. API efficiency and understanding and performance are also key on our list to actually improve. Cross-wicky features such as notifications and search, being able to explain what we can see and where we can see it. Data and feedback tools, such as quick survey, and some of the data APIs that we're going to be able to continue to improve upon.

8:32And bringing O-RES, translation, graphs, maps, Wikidata Query Service, UI consistency, and making sure these types of things actually get built into more and more features. The engagement and prioritization component of this is really us as a foundation doing a better job of making it clear what we're working on why we're working on it and when we're going to be working on it one of those things that i brought up earlier was a community wishless survey we had over a hundred uh community um suggestions put forth these are the top 10 uh there's going to be a discussion tomorrow i believe uh worth a session where we'll go more in depth about how that worked and where we're going to go with it i think that this is a good example of how we're We're trying to do a better job of actually engaging with people, getting the right things, and building the right things.

9:31So working with foundation teams, the primary way to work with us is really through our team fabricator boards. Making suggestions, engaging in the discussions that are ongoing constantly, just like we did for the developer summit, are the primary way that we would like you to engage with us at this point. Participating in planning really comes from engaging in the team wiki pages, the mailing lists, the talk pages, IRC, and our fabricator boards. Monitoring the progress of our goals and our projects comes from going to our, we have set up our goal pages on MediaWiki and we're trying to do a much better job of putting those up early and announcing them on our mailing list and engaging people on a more regular cadence so that there's more warning and there's more participation and discussion that can occur.

10:26We've also started an initiative around product showcases this year, where we're actually trying to show the work that we're developing so that you have a much more visible way to interact with us. And then lastly, we've also created a product hub where you can see the status, you can see metrics and outcomes as we perform, and there's a number of jump points within there to go to all of these pages for each product team. Really, more communication is the whole key. We want more blogs. We want to continue to work with developer relations and community liaisons to talk about the things that we're developing. We're looking at our development process, and we also are trying to engage by saying when we're running user tests, what surveys we're running, the results that come from them, and have a discussion around our strategy.

11:12The teams at the foundation are broken into technology groups and into a product group. And I just provided these really as jump points to be able to go deeper into those particular areas and teams that you can work with. And lastly, I just wanted to share some of the milestones that we have currently going forward. The goals for this current quarter are posted through January and March on MediaWiki.

11:46you can engage with each of the teams and ask them directly how how they came up with it what the rationale was around it but we want like I said earlier to do a better job of doing an earlier cadence on these sorts of things so we want to try to have our planning and drafts for The following quarter for April through June done in February and March and ideally we want to get to a 9 to 12 month road map where you're going to have a much longer window to actually engage with us. The planning and review for a lot of the work that we're doing at the foundation is going to be happening around the annual plan for 2016 and 2017 from now until June and I highly encourage you to engage with us.

12:29So, I want to keep it relatively short and provide an opportunity for questions and answers. But I wanted to make sure that we had a broad picture of kind of where we've gone, where we're going. This event, the annual plan, and ongoing conversations directly with the teams as they develop their goals is the best way to engage with us right now. And, um, so thank you for today.

13:01Thank you for the discussions that we're going to have. And I look forward to participating with all of you. Thank you. Thanks, sir. And one of the things I would love to celebrate is the, is back on the community wish list side of things. We've, a lot of folks put in a lot of work here and I want to make sure that everybody knows the people because like one of the things that we're hoping happens here at the summit is that you talk to these people about what the community wish list items were.

13:39So, Danny, yes, so Danny Horne here, and Danny, if you don't mind me putting you on the spot, I'm going to let you introduce the rest of the team. Hi, everybody. I'm Danny. I'm the product manager for the community tech team. We also have Ryan Kaldari, our technical lead, Niharika Kohli, and Francis Hoga. Oh, she's over there. There's Francis. Yeah, we would absolutely love to talk to everybody about, we're basically starting 10 huge projects at the same time.

14:17There's a lot of conversations we need to have, so that will be great. I will talk to you all soon. Great, yeah, I think a lot of the projects that you see here, these are all things that I know... We've, you know, we've been talking about a lot of these things for a long time and talking, but not really like having a good concrete plan for them and so I love the fact that the community team has been working on actually figuring out how we move forward on these things and starting with, you know, the first step on these this year, I mean 10 projects, that's a huge amount of work for a very small team to do, but with the help of everybody here, we can get it done. And that's, and not just everybody here, but everybody in the world that is a programmer. Like, we want to get those folks.

15:10Kim, that's also something that's going to be highlighted at the hackathon as well, right? Opportunities to work on this list. So, cool. So, anybody want to comment on any of these projects? Like, that would be a good start. Like, is there, is there anything that strikes anybody here as like, like, oh my God, we, if we do one thing, this is the one thing we have to do. Yeah. Okay. Brian. I think for number five, there's an option we could just enable in the underlying library.

15:46So it's just a simple matter of configuration. There's a global, you said. It's fine. So, yeah, I think that having that. So at this point, there's starting to be people. I'm breaking with our plan here by holding the microphone and being microphone guy and letting people sort of sit in their seats and do this. So if you can line up at the microphones on either side and we'll, and we will take people in order, in their order in line. So, yes.

16:26I guess I kind of had an unfair advantage because I was sitting literally right in front of the microphone. But considering the broad scope of the product department, what would you say is the underlying mission of the product department? Or in more operational terms, how do you decide what to work on and what not to work on? So each individual audience right now definitely has the responsibility to take in a number of different feedback loops to figure out their prioritization. As a larger product department, we're doing some strategy development as part of the annual planning where we're going to open it up and have active discussions around it.

17:04This is new. It's relatively new. I started just technically two months ago. But I've been at the Foundation longer than that. We're really trying to drive people to the teams themselves because they're closest to what's going on. They understand the audiences and so, like I said, try to work through their fabricator boards, try to work with their mailing lists, try to work with their conversations as they build their goals and their planning. Does that answer? Yep.

17:43All right, so on the community wish list survey, I just want to comment. I think it's all about number three, Central Global Repository for Templates. Is this on? He said it's all about number three. Okay, good. Talk into it, yeah. So, ooh, that's better. So yes, I just wanted to add that many years ago we set up Commons for Central Repository for Images and it made it hugely easier to reuse things between projects and between languages.

18:21And it's really, really, really been missing for templates, modules, and gadgets. It creates an extra complication for authors and maintainers. And there's no good localization. And so there's been a lot of. Difficulty in terms of making these materials actually available for use on what is now several hundred separate wikis. So, in my opinion, that is the most important thing we can possibly do and I think a lot of people will get benefit from it.

18:53Thanks. I'll jump in really quickly to say that about four years ago, Timo and I wrote some code that did that for gadgets and never finished it. It is my understanding that the community tech team has already begun picking that up and polishing it and doing things to it cool to that I would add another indication of how important that is is I think nine out of ten of these can be implemented using a gadget which is not necessarily advisable but it kind of illustrates how important having a central maintainable repository for that would be if you're interested in talking about number three I have an RFC called shadow namespaces that would take care of templates and Lua modules I'm gonna schedule in for an unconference slot if you want to talk about that all right that's 20 minutes thank you

Source and licence

Recording: File:DevSummit2016-Opening-Keynote-by-Wes-Moran.webm by Wes Moran (2016-01-04), via Wikimedia Commons. Licence: CC BY-SA 4.0. The transcript, summary and chapters on this page are shared under the same licence (CC BY-SA 4.0).

The transcript is machine speech-to-text by the TidyTools Audio & video transcriber; the summary, key points and chapters were written by Claude from that transcript. Both may contain errors: check the recording before quoting it.