MIskander-WMF

SpBot archives all sections tagged with {{Section resolved|1=~~~~}} after 3 days and sections whose most recent comment is older than 90 days.

Technical debt

Hi!

The WMF should hire more technical people to work on Phabricator tickets. Can we pause shiny new projects, and prioritize working on existing problems and technical debt instead? Any website that exists for literal decades(!) builds up technical debt over time, and it is good to focus on that once in a while. Polygnotus (talk) 23:41, 18 March 2025 (UTC)

@Polygnotus: I saw this message and wanted to respond as my teams look after core MediaWiki functionality. I agree, technical debt is important. It is why the organisation's annual plan treats sustaining our infrastructure as essential work. The Product and Technology teams dedicate a permanent, year-round priority to the maintenance, accessibility and operation of the wiki projects. Last year Foundation staff were able to resolve 1500 tasks opened by community members (see tech news). I can see on Phabricator you have been having issues with massmessage - a tool that is showing its age. The Content Transform team in our Product & tech dept can look into this further and will respond on the Phab ticket. (They are Content-Transform-Team on Phabricator.)  My colleague @ELappen (WMF): can help with your other question. SCherukuwada (WMF) (talk) 16:19, 21 March 2025 (UTC)

Thank you! I wouldn't go to the CEO of the WMF to get help for just my Phab ticket. I want to ask the WMF to hire more technical staff. I think that there is broad support for this wish. Resolving 1500 community-reported tasks in a year is a great step in the right direction, so let's scale that up. Polygnotus (talk) 17:33, 25 March 2025 (UTC)

Update: I looked at Phabricator, and it seems to me that a lot of time is spent on WMF pet projects. There appear to be are a large amount of tickets that serve mostly to benefit people who work at the WMF, and not the community. And it looks like those tickets receive more attention. How can we solve that problem? Who decides what tickets are important and which tickets are not so important? If there isn't a robust system in place to determine what should be prioritized (I can set someone else's task on Phabricator on low priority and my own on high priority) then the resources are not utilized optimally.

Focusing on specific examples can be unhelpful, but it may be wise to highlight something like Special:AllEvents (aka the Collaboration list, part of CampaignEvents), which is something I didn't even know existed and apparently cost a very significant amount of devtime but offers basically no benefit to Wikipedians. There are at least a dozen of other examples. Polygnotus (talk) 18:01, 2 April 2025 (UTC)

Hello @Polygnotus -- I'm Marshall Miller; I'm a director of product at WMF and @SCherukuwada (WMF)'s colleague, focused on planning our priorities.  Though almost all of our work is tracked in Phabricator at a granular level, it's difficult to see on Phabricator how the many tasks all tie together into larger projects.  I want to fill you in on how we decide what to prioritize.  When we make our plans, we are thinking about the long-term needs of all different sorts of users of the Wikimedia projects.  There are many different kinds of editors doing different things in different languages, plus the needs of millions of diverse readers.  Though it may seem like we have a large team, it is actually quite small in comparison to teams running other large websites, and so we are constantly balancing and making trade-offs across all these different needs.

To figure out what to prioritize, we look at global trends around how the internet is changing, as well as internal trends around our editors' needs, which we learn from data, from Phabricator tickets, from the Community Wishlist, from the PTAC, and from many conversations on-wiki and at conferences.  We then decide to take on some large projects that we think are important investments for the future and that only the WMF is well-placed to tackle (i.e. the sort of projects that it would be hard for volunteers to do on their own).  The case that you mentioned, of the CampaignEvents extension and its Collaboration List, is a major priority to volunteers who organize editing events on-wiki and in-person, helping them gather first-time volunteers and generating substantial content.  And in the long term, we plan to extend it to help volunteers work together on-wiki -- for instance, the Collaboration List will become a place for new editors to find WikiProjects to get involved with, so they can connect with others and find a place to belong on the wiki -- and then hopefully stick around for the long term and become the admins and patrollers of the future.

Another example of a major project is called Edit Check, which guides new editors to make edits that comply with wiki policies.  This gives them a better experience, and also cuts down on the number of unconstructive edits that patrollers have to revert.  We decided on this project after hearing different versions of the idea from volunteers in the Community Wishlist, and after realizing that improving machine learning technology would make the idea possible -- it was an opportunity in which a need of experienced editors aligned with the needs of newcomers, and was enabled by a global technology trend.  These major projects turn into many Phabricator tasks as the teams implement them.

To sum up, we determine a set of these large projects that are priorities.  And then, as Suman said, we reserve substantial time for maintenance and tech debt, which we call "essential work".  You can see how our planning process works and what it contains here on Meta.  Here's where you can look to see the plans that we're executing on now, and here's where you can see the coming year's plans being built (note that our "year" runs from July to June).

Please check this material out and let me know what you think -- I am happy to answer questions and hear your thoughts around other projects.  You can contact me on my talk page or at mmiller@wikimedia.org. It's important that as many people as possible understand why we're prioritizing what we are. MMiller (WMF) (talk) 18:50, 4 April 2025 (UTC)

@MMiller (WMF) Problem is, WikiProjects are not actually how Wikipedians collaborate. You'd think they do, and it would be reasonable for anyone to expect that they do, but they don't.

In reality some people have watchlisted every page that is of interest to them and their WikiProject, and if a new user shows up and makes some edits they'll get to know them soon enough (either because they revert imperfect contributions or thank them for good contributions. And people don't really need WikiProjects, because most people just remember a few users and what they're good at. When they have a question or idea, they'll talk to whoever they know has expertise on that topic. If they don't know anyone they might use a noticeboard.

Getting new users to join a WikiProject makes no sense when most WikiProjects are lifeless and (almost) pointless (there are rare exceptions, but those are so rare that they don't justify all this time and effort and money).

WikiProjects lack the one piece of functionality that is actually important (and yes I've thought about writing a script before, but using a bot would be better):

• Sort the list of participants in a WikiProject based on recency of their last edit and the amount of edits they've made (with a weighted score system)

Even on the English Wikipedia, with the most active users, WikiProjects are basically vestigial. So for the WMF to spend a lot of time working on that stuff is unwise.

Same thing with the CampaignEvents and Collaboration List. You call it a a major priority to volunteers who organize editing events on-wiki and in-person and that is probably true, but almost none of our users participate in in-person editing events, and almost none of our new users come from in-person editing events. So while it may be a major priority for them, that is a major priority of a tiny tiny group of people.

It looks like the WMF has people who focus on that stuff, and somehow they think that it is a great idea to spend a lot of money to develop tools to make that easier/better et cetera. But in reality this does nothing to help the existing community, and the tiny amount of newcomers it generates (at huge cost) usually don't stick around, or get scared off when they are bitten once or twice.

I am nothing if not bold, so I added one of my phab tickets to a Kanban board of the relevant team. The response was: It is up to teams to decide what they plan to work on or not. If that is true then that explains that people focus on building sleek new interfaces and bolting them on top of the dinosaur codebase. What any company needs is someone who decides what to do and, perhaps more importantly, what not to do who can look at the big picture and make unpopular choices. A slightly adversarial role to keep the nerds from doing whatever they like. The teams should not decide what they work on, because they have a vested interest in working on stuff that is fun (creating a shiny new interface, doing something with machine learning/AI) instead of what actually needs to be done (rewriting boring 20 year old code or replacing it, dealing with boring Phab tickets). And the teamleader has a vested interest in being liked by their team. Am I making sense?

The "be evil" company known as Google gives their teams 10% of their time to work on whatever they want, and 90% of the time they work on stuff that is actually important according to someone who can see the big picture and knows what to prioritize.

You don't have to believe me without evidence, go look at the viewcounts of the WikiProjects of the English Wikipedia and how often they get edited. Most people put their name on the list and forget about them. Each WikiProject is dominated by at most 3 users who try to make the rules for everyone else (which can be a good or bad thing, but is usually a mix of both). Ask around what the return on investment is from in-person editing events.

The Edit Check thing is a good idea, generally speaking, but you guys are over-engineering it to hell and back. You need to figure out where to stop. When something is "good enough".

Polygnotus (talk) 20:19, 4 April 2025 (UTC)

@MMiller (WMF) Do you have people who can tell you which ideas are bad? Grumpy people, with a different communication style? My impression of the WMF are that you are all such nice people, who would never contradict each other for fear of offending them, and hide all honesty in a compliment sandwich.

I have no clue what "director of product" means, probably nothing, but if you need help determining if something is a bad idea you can just contact me.

I skimmed through Wikimedia Foundation Annual Plan/2024-2025/Product & Technology OKRs and now I know less than I did before I read it. It appears to be a list of vague platitudes and truisms. The sentence Invest in large new features like chatbots or social video on our platform? is baffling. See how they want to use an LLM to detect peacockery, but they do not want to rewrite 20 year old PHP? I have exactly the same problem, every day. And if I didn't have a boss (wife) and a "manager" I would play around all day with my favourite toys and the important stuff would not get done. Polygnotus (talk) 00:48, 5 April 2025 (UTC)

@Polygnotus We appreciate all of these insights and perspectives. It is clear that people can have different views about how to prioritize what is most important to which groups of volunteers. That said, personally insulting Foundation staff in the process is not acceptable and being civil to each other is the only way we can engage further. Thank you for being in touch with your questions and contributions. MIskander-WMF (talk) 14:08, 7 April 2025 (UTC)

@MIskander-WMF What insult are you talking about? I have been nothing but polite and helpful. We appreciate all of these insights and perspectives. I don't get that impression at all. Falsely accusing someone of incivility is, of course, incivil. So please explain what you mean. Polygnotus (talk) 14:12, 7 April 2025 (UTC)

I just re-read this page, and nothing I said is in any way a "personal insult" to anyone, including Foundation staff. Casting aspersions without evidence to back them up is not just incivil but also a terrible strategy. I expect an apology, and a retraction of your false accusation. Polygnotus (talk) 14:20, 7 April 2025 (UTC)

The only thing I wrote that anyone can object to is that I joked about Google, but that is simply an old joke about the fact that they dropped the "Don't Be Evil" motto.[1] Is that what you mean? Polygnotus (talk) 15:47, 7 April 2025 (UTC)

Hi again, you made a reference to having no clue what "director of product" means, probably nothing, but if you need help determining if something is a bad idea you can just contact me. You have my apology if I misunderstood your intent in this response, I don't plan to engage further here. MIskander-WMF (talk) 16:28, 7 April 2025 (UTC)

Are you saying that the fact that I am not familiar with every single phrase in the English language is insulting to you, or the fact that I offer to help weed out bad ideas?

It seems like you admit that there is no personal insult to foundation staff and then you post a conditional non-apology. Are you sure that is the best approach?

I am not a native speaker, and I don't think I've ever heard/read that phrase before. This is a worldwide community, and while my English is better than most, it is unreasonable to expect everyone to know everything. I don't expect you to speak my native language.

I do know what "CEO" means, and I do know that this is a very bad look.

How can you claim to misunderstand my intent when there is no reasonable alternative interpretation of my words?

You falsely accused me after I posted valid criticism (in a polite and constructive way). I know correlation does not necessarily imply causation, but not everything on this planet is random. I still expect a retraction and an apology.

Writing you don't plan to engage further is quite insulting in this context. I helped you by politely pointing out a problem, you falsely accused me, I asked for proof, you were forced to admit there was no proof and had to backtrack. So I think you should engage further; I am specifically interested in your response to my point that because WMF employees do whatever they feel like a lot of the important but boring work does not get done. That is information that should be very important to a CEO, especially in the context of what I described as a "dinosaur" codebase. Polygnotus (talk) 07:32, 8 April 2025 (UTC)

Before I forget I should add that I am also very unimpressed by the fact that you post a compliment sandwich in response to a comment that complains about compliment sandwiches.

I interpret such attempts at manipulation as an insult to my intelligence, and I worry about people who speak their truth only 33% of the time. That can't be healthy.

I've been told that this pattern (falsely accusing those who try to help of incivility, even if they politely offer constructive criticism, and then refusing to communicate when the accusation is proven to be false) is a trick used by people within the WMF. Polygnotus (talk) 18:29, 11 April 2025 (UTC)

Asked Claude to write a script. Inactive means no contributions for more than 1 year, blocked means indefinitely blocked.

Wikiproject Trains: 253 active, 246 inactive, and 19 blocked users.

WikiProject Buses: 43 active, 52 inactive, and 2 blocked users.

WikiProject Transport: 54 active, 54 inactive, and 2 blocked users.

Wikipedia:WikiProject Aviation: 312 active, 172 inactive, and 11 blocked users.

WikiProject Geography: 45 active, 41 inactive, and 7 blocked users.

So roughly 50% of those who are on a list of participants of a wikiproject are inactive.

And I found 205 Wikiprojects in en:Category:Active WikiProjects that haven't been edited at all in a year https://en.wikipedia.org/wiki/User:Polygnotus/inactivewikiprojects

Polygnotus (talk) 16:08, 6 April 2025 (UTC)

Organizational chart

Looking at the Wikimedia Foundation/Organizational chart, it appears to be outdated. Some of the people have left (e.g. User:GBordoy (WMF) and User:AParker (WMF) and User:GBasha-WMF) and others never had a WMF account created (e.g. User:BKurgat (WMF) and User:AEspinoza (WMF)). It also appear to be incomplete (en:Wikimedia Foundation claims a total of 700 staff + contractors). For the average Wikipedian it is really difficult to figure out which WMF-er to talk to about which topic. Would you be so kind to ask someone to update the chart? Which departments does the WMF have and how many people work in each? Thanks, Polygnotus (talk) 00:02, 19 March 2025 (UTC)

Hi @Polygnotus: thanks for the flag on the org chart. In terms of your need to reach someone in particular at the Foundation, we’ve done work on the Foundation pages and processes to make it easier for community members to get connected to the right staffer. We actually want to take the burden off you of combing through an org chart to try to guess who the right person might be, and so have set up an entire triage system to support community questions and requests. If you visit the Contact us page (or click on the button in the top right corner of all Foundation pages here on Meta-wiki), you’ll see information about how to reach out to us. We’ll then triage your question, making sure to connect you with the right person within a few business days. This helps keep it more light-touch for you, and also ensures that the right person sees your message and replies.

In terms of the org chart page itself, I’ve added a disclaimer to the top of the page to clarify the cadence at which the page is updated, and added a pointer to Contact us. Our org chart is not intended to be exhaustive, but rather a directory of leadership, as well as public-facing and community-facing staff, to give some faces to the work. Product & Tech is a bit sparse currently as we’re working to get these new staff listing templates on mediawiki.org and start using them there–that is coming. However, as an overall approach, to respect staff privacy and in some cases to comply with local law, we are committing to listing people in these roles plus those who have requested to be listed. This helps us balance privacy and transparency, while our improved Contact us system ensures practical needs are met. --ELappen (WMF) (talk) 19:42, 21 March 2025 (UTC)

Thanks! Leaving a note on the org chart is a good idea. Which departments does the WMF have and roughly how many people (or FTE, if that is a better metric) work in each? Polygnotus (talk) 17:18, 25 March 2025 (UTC)

The departments are all listed both in the Foundation’s header on Meta (which appears on all official Foundation pages, for example this one)--you can click to drill down into the teams. Departments are also listed in the org chart page you’ve referenced. 57% of the organization’s headcount, as well as close to 50% of the organization’s budget, is dedicated to our Infrastructure Annual Plan goal, which is Product and Technology work; this reflects our growing investment in infrastructure year over year. That linked page has more information about headcount and budget dedicated to the other goals as well, in case you’re interested. I’ll also point you toward our recent Annual Plan update, which details the work and investments made in Q1 and Q2 of this fiscal year (July-December)--it gives lots of information about the Product and Technology work that’s happening now. ELappen (WMF) (talk) 19:27, 25 March 2025 (UTC)

@ELappen (WMF) Thank you! So if I understand you correctly the answer to my questions is that the WMF agrees with me and is moving in that direction. That is good news. Polygnotus (talk) 20:51, 25 March 2025 (UTC)

@Polygnotus joining the conversation to say yes, we are all agreed on this priority. Thank you for reaching out. –MIskander-WMF (talk) 10:27, 26 March 2025 (UTC)