Hey everyone!
Not too long ago, Fandom Stars sat down with Fandom’s Vice President of Engineering, Daniel “Grunny” Grunwell, for a Fireside chat on Discord. Stars gained insights into the engineering team's vision and priorities around the editor experience and product roadmap. They also learned how feedback directly influences the platform's development.
For those unable to attend, this recap highlights the critical questions from Stars and information from the Grunny, ensuring that the entire Fandom community benefits from the insights and discussions that emerged during this special event. Stay tuned for updates and further opportunities to connect with members of Fandom’s staff and contribute to shaping the platform that continues to empower and inspire fans worldwide.
Key Questions and Answers from the Roundtable:[]
Tell us about yourself?
Grunny: Sure, hey folks! I'm Grunny. I have worked at Fandom for over thirteen and a half years now, but I started editing on Wookieepedia in 2005, back when Fandom was still Wikicities. I am still an administrator there (formerly a bureaucrat), so I've been editing on wikis for over 19 years, which makes me feel old to say.
I'm a VP of Engineering here at Fandom where I support the various teams working on the User Experience for the wiki platform; user experience for GameSpot, TV Guide, Metacritic, GameFAQs, etc.; the Wiki Platform and Services Platform teams; the Engagement team (who work on features like Interactive Maps); the User Generated Content team (focused on editor and admin tools); the mobile apps team; and the Fanatical tech team.
What are your fandoms?
Grunny: I have too many Fandoms, and I have a tendency to get obsessed with yet another one regularly, but my main ones are Star Wars with Star Trek as a second, most things cyberpunk and sci-fi generally, Teenage Mutant Ninja Turtles, and various anime. More recently my favourite book series has been the Murderbot Diaries for the last couple of years, and I'm once again rewatching Babylon 5. I also obsess over karate and tennis.
Where does the engineering team fit into the broader Fandom ecosystem? Also, how do you think engineering could benefit from greater interaction with users and vice versa?
Grunny: Engineering is the biggest team at Fandom, after all it takes a lot of folks to help build, maintain, and scale a platform that is serving hundreds of millions of users every month, and making sure everything is running and working while then trying to build new features and solve problems for users.
We are tightly part of the "Tech and Product" group as engineering, product, and design are all a tight partnership to figure out how we continue to both provide extra value and keep the platform running well at such large traffic. We then are also closely partnering with the Community team to conenct with all of you, marketing, sales, and other folks all across the business.
How do you work with the Community Team?
Grunny: It's a great question. For our product teams we structure each area as a "triad" for tech (engineering), product, and partners like community. Each area has a community team member assigned to it who is involved in all releases, bug triage, and planning. So, each product area has someone like Antonio or Lostris working with them directly on a weekly basis so they can help with the community perspective and communicating with all of you. The product lead, engineering lead, and community lead collaborate together on the product plan, how we test with communities, gather feedback from all of you, and communicate new releases.
Does the team check the platform's viability automatically or does it rely on user reports to know about bugs and broken features?
Grunny: It's both. We have a lot of automated checking. For example, a large test suite is run before every release to try to prevent any issues that would break key functionality from being released to users; we have monitoring on page performance, errors, and more for various features and services so that we get automatically alerted if something starts misbehaving, going slower, or causing a poor experience. We always strive to notice those things before it impacts our users, and we do most of the time so we can react to address the issue and ensure user impact was minimized (our team is constantly checking graphs and dashboards to make sure things are working well), but sometimes things slip through, and we get the reports from all of you on issues and bugs of various types, given both the very large platform, large number of features, and wikis with various customizations, there are a lot of edge cases we need to hear about so we can make a better experience.
How do big errors slip through with such a mechanism?
Grunny: There are lots of ways errors can slip through in a very complex/high-scale system, even despite best efforts, but what we do as well is always try to figure out how anything that did slip through happened and how we can avoid it next time. For every incident, the team goes through a root cause analysis process (RCA) where we work through an honest look at what happened, the timeline, keep iterating on questions until we have the root cause identified, and then a plan to try to prevent it happening again so that we're constantly improving, adding additional monitoring/alerting, additional tests, adding things to MediaWiki to make it more resilient to certain things, etc. so that hopefully we're constantly improving.
The conversation around AI has changed a lot from even last year when it was discussed at Connect 2023, how do you think AI fits into what Fandom does on the technical side, and also on the wiki side?
Grunny: I think we're exploring where it makes the most sense in various ways, such as how it can be a helper to make certain jobs easier. On the technical side, we're exploring how it could help improve testing for example, by identifying anomalies in automatic tests of the platform, or in helping engineers create more tests. On the wiki side, we're going to be exploring various ways it might help editors with some things, for example, could it help flag an edit by a user for review by admins quicker as potentially problematic/vandalism, as GenAI is able to, in theory, understand more of the surrounding context if trained well. But we have to see how those things work as tests and if they add enough value to the experience, so ideas of course, are welcome too.
Will it be possible for more wikis to enable Semantic MediaWiki or Cargo in the future?
Grunny: For Semantic MediaWiki and Cargo, no. Those features have a number of drawbacks that make them problematic at scale. They are also very user-unfriendly to new users/editors (or even experienced editors, let's be honest). So, related to my previous points, we want to learn more about some of the use cases on SMW/Cargo that might be the most valuable and see how we could provide that or similar functionality in a user-friendly way, and a way that doesn't cause performance or other issues.
Are you exploring new ways to implement Structured Content in the platform?
Grunny: Yep, we're actually in brainstorming and planning mode on this at the moment for next year. In particular, we're thinking through the types of things editors are already "structuring" to be easily understood and reusable and wikis that are doing this in interesting ways so we can figure out how we could make the editing experience easier.
Are you planning to upgrade MediaWiki again soon? And how it will impact the wikis?
Grunny: We're currently in early stages of planning the next MW upgrade. We need to continue upgrading, but we're trying to find the right balance of timing since we know it impacts all of the wikis a little as the upgrades roll out and might require various small updates to CSS and other things. So, we're currently planning to figure out the exact timing that will start in the coming year, and working with the community and product teams on the right timing as well as to be able to communicate the plan to all of you, including with details of the changes in the new MW version we'll be upgrading to.
Where do you think Fandom has improved the most from when you first started?
Grunny: This is a great question, and it's been such a long time it's hard to pick just one thing. Since I first started editing, the platform has changed dramatically. It's obviously a lot more usable and provides a lot more features for editors and readers. I actually have a screenshot of the Fandom Community Central equivalent page from the time I created my account if you all are interested in what it looked like!
I think we've come a very long way. When I started, we had wikis in the hundreds to early thousands (wiki creation at the time required a request for the wiki; you couldn't create your own), and we've scaled to 300M users a month, 200K+ wikis at any given time, and lots of extra features. in terms of when I started as an engineer at Fandom, we've also come a very long way in terms of platform scale and stability, I remember when every couple of days back in 2011 an engineer might accidentally cause a "white page of death" with a rollout, and getting paged with an outage was a regular thing, where now it's very rare for such an issue to happen, and we have one of the quietest on-call rotations I know of, as we have a lot of people working on making this a very stable and reliable platform and reacting proactively to issues (always more to improve, of course, but it's something I think we do well)
Check out this screenshot from Grunny’s Wikicities account and apply for Fandom Stars if you want to join similar conversations and provide productive feedback on the platform!