User:Clovermoss/Mobile editing/Talk page discussion

Software development challenges

I'm following up on your comment here, because it's not directly related to your main point (I haven't looked into the details of spending, but I'm certain software development can be done more effectively, because that's almost universally true). I can't comment much on the WMF's software project management, since I haven't followed it (though from what little I've seen regarding its interactions with the community, I do feel it can be improved). There are a lot of challenges to it, though. The MediaWiki developers strive to support a broad set of devices with different levels of technology and Internet connectivity, and users with different levels of knowledge and limitations. Every user has their own set of priorities on what they would like developed, and it's hard to figure out what features would satisfy the most people. Free software lacks the "will users pay for this?" filter to drive feature development. This isn't all bad, but it means there's less urgency to figure out future plans and to complete features.

Taking talk pages as an example: they use the same wikitext markup as in articles (as I understand it, discussion used to just be held at the bottom of articles). Some editors are adamant about preserving this, so there's no difference in editing an article or a discussion. Many editors were understandably displeased by the WMF's attempt to develop an infinite-scrolling, threaded discussion forum interface, and are now very wary of changes. The talk pages project is trying to navigate a middle ground where incremental changes are layered on top of the wikitext conventions.

When seeking input on MediaWiki on Wikipedia, you get feedback from a self-selected set of interested users, and there's no guarantee you'll get be able to get prolonged engagement from them to provide ongoing feedback. You can use telemetry (monitoring of user actions) to gain insight, and from things that WhatamIdoing have posted, it seems the development team looks at feature usage and editing patterns. You can use focus groups, but with a community as expansive and anonymous as Wikipedia's, it's hard to get a representative sample. There's also a significant portion of the community who think the WMF should provide the bare essential functions, and let the community develop everything else. Recall Levivich expressed his desire for a third-party Wikipedia editor to be developed (though as a competitor to the in-house provided editing interfaces). Numerous other editors think all visual editor development is a waste of time.

In summary, I agree the WMF should try to do more to overcome the challenges it faces, but I acknowledge that the challenges are daunting and limit what can be done. isaacl (talk) 04:00, 30 June 2022 (UTC)

I see you've put a lot of thought into this. I wouldn't say I disagree with the heart of what you're saying. There's a lot of nuance. I think one of the reasons everything is complicated is because everyone involved is super passionate and this can lead to kind of seeing the trees instead of the forest, y'know? Or not recognizing that certain plans are flawed until there's a larger outreach. I do think better communication and awareness between the WMF/the community at large is a good idea, as I previously stated in the growth team discussion at Levivich's talk page. There's also optimism bias. I'm a fairly optimistic person myself, but it's good to be realistic about certain things. Change doesn't come from nowhere, but thinking you're the only person capable of changing something and that you'll suceed where everyone else has failed often doesn't lead to a good outcome, either. Because there's been some bad precedents (or at least what has been percieved that way), there's a lack of trust, too. Ideally, the overlap between the WMF/broader community shouldn't be comparable to a venn diagram. Of course you're more likely to hear from the people who are upset with certain changes, but again, I don't think ignoring something that's reiterated by mutiple people is the best solution to that. Again, you obviously can't please everyone, but people in general shouldn't feel like their concerns are being ignored.

There's also that the very nature of requesting feedback from people you're more likely to get people who are super involved/passionate and that by itself is a form of survivorship bias. But at the same time, these super involved people care. Sometimes these concerns can be dismissed (and this happens a lot in real life too) as "haters being haters". But I think it's important that there are people who aren't afraid to express how they feel. I'm someone who has a hard time standing up for myself in general, but I feel like that's more understandable because of certain adverse life experiences. But just because someone's rude doesn't mean that there isn't any point to what they're saying. And just because someone's nice doesn't mean that they don't have misgivings. That's not to say I don't think being civil is important, because it truly is, but sometimes being straight to the point and trying to call things out is nessecary and there's not really an ideal way to do that without emphasizing that the way things have turned out have left you frustrated. Again, not saying that this makes it okay to attack people, but again, I'm saying that there's nuance to everything involved. I heard a radio ad once that was along the lines of "there's homeless teens... here's some statistics about how much their life sucks" and then it cuts to someone swearing. The point the ad made was that people were more likely to be upset about the swearing than they were about the homeless teens. Civility shouldn't be just a pretense, but it also shouldn't be something that is somehow more important than vital issues. Not that whatever goes on at Wikipedia is on quite the same level as life-changing in most circumstances, but there are a lot of potential real-world ramifications of what goes on around here. The easiest example of this is with BLPs, but there's more than that to be concerned about when it comes to the impact Wikipedia can have to the world at-large.

In general, I don't think people that criticize the WMF are saying these things because they want the WMF to fail, but such input may be ignored regardless. I think it's reasonable that people get frustrated with that and that that leaves a lasting impact. That people keep talking about whatever happened for ages. In general, trust is something that's hard to regain. Especially if someone feels like it's been broken over and over again.

You also don't hear from all the people who quit in the process. Or get to listen to their perspectives. At least not typically. I read something once a few years ago that stuck with me. It was a comment along the lines of "we need more people who don't make a career of editing the encyclopedia." It's one of the reasons I thought that my idea of an expanded #1Lib1Ref but #1Person1Ref would be useful. The potential impact of "true" crowdsourcing has a massive scale. Trying to basically compete for the attention of active editors for growing backlogs can only do so much. We need more editor retention, but we also need more editors in general. But getting more people involved with a change like learning how to cite a reference, which is a small but achievable goal to teach people, could get people more interested in Wikipedia editing in general. Obviously there's going to be a ton of people who never do anything of the sort again and move on with their lives, but I think to a certain extent Wikipedia editors are self-selecting. A little taste of what editing looks like could prompt something more, even if it's just a fraction of a fraction of an initative like that. But isn't that how things already are? A fraction of a fraction of people who create an account make an edit and then continue to edit, let alone be what most people would consider an active member of the community.

By the way, I think this is the longest talk page comment I've ever written. I knew my desire to write walls of text would blossom one day! Joking aside, I'm quite talkative in real life, so this isn't really all that surprising to me. I've got to say that reading this back to myself it seems like I've made some really good points but it's 3 am so maybe not all of my ideas are formed out the way I would want to be. I'm going to go sleep and maybe I'll have something else to say later on once I've had some time to reflect and not be exhausted. Clovermoss (talk) 07:21, 30 June 2022 (UTC)

So after going back to sleep, I wanted to add a few things. First is that I agree with basically everything you were saying but I understand aspects of other people's viewpoints. I think it's important to have conversations where people disagree with each other, because reasonable people can disagree. Ideally, everyone is civil and no one attacks each other. But there are ways to have good intentions and nice, but also cruel. This is more of a general life thing. A lot of what I was saying has implications beyond just the way the WMF/the local community interact with each other. I think more effective communication and more participation from the broader community is an important thing.

I think the main reason nuance is so important to me is because I grew up in what I would call a restrictive religion. There's a million things you can't do if you're in "The Truth". Reading the article about Jehovah's Witnesses would likely be useful background for what I mean. Anyways, when I was about 5 or so, one of my parents was disfellowshipped. Literally no one in my family who was a part of the religion has talked to them since. Not their twin, not their mother, etc. People would act as they were dead, to be honest. Whenever people would talk about them, it was in the past tense. They "used to be such a good person", "it's such a shame they're no longer with us", etc. But it was all from a place of good intentions. By shunning them, they were showing that they truly loved God and wanted my parent to return to the Truth. I grew up with people warning me to not turn out like them and that they were being influenced by Satan. All because they had a different idea of what it meant to be Christian. But that was apostasy and apostates were "mentally diseased"... there was a few times I had literal panic attacks that hospitalized me as a child at assemblies when they'd talk about apostates because I knew my parent was one of them.

It's hard to explain where I'm coming from if I don't explain some of the things that made me the person I am today. So sometimes I focus on things that people take for granted, I guess? Like seeing the nuance in everything is a very important thing to me. Clovermoss (talk) 11:36, 30 June 2022 (UTC)

@Levivich: Pinging you in case what I wrote above changes your position about what was said at COIN/Lamb v Benoit. It's more complicated than just being a former member of a religion. Clovermoss (talk) 14:31, 1 July 2022 (UTC)

<not-your-dad mode> You know these message are timestamped, and four hours of sleep is not enough sleep. I hope you turn in at a reasonable hour tonight! :-D </not-your-dad mode>

It's still not a COI, unless those apostates are paying you for your apostasy. :-) A bias, sure, but we can edit articles we're biased against and still edit them neutrally. Look, I edited Hitler Has Only Got One Ball because I hate Hitler and want this notable example of mockery of Hitler to be known to the whole world. My bias against Hitler is what motivated my volunteerism in this case; still, the article is neutral. Same for George Floyd and most everything I work on. Or for example Elgin Baylor -- I think he was an underappreciated basketball player, that's why I worked on his article. I'm biased in his favor; doesn't mean I can't edit neutrally.

You're perfectly aware of your own bias on this topic, which is 90% of what's needed to edit neutrally. The other 10% is a good night's sleep :-) Levivich^[block] 16:40, 1 July 2022 (UTC)

@Levivich: I woke up in the middle of the night and wrote the original comment, so I technically got more than 4 hours of sleep. In general I do get a decent amount of sleep but sometimes I work day shifts or night shifts and that can kind of throw everything out of whack. Also you're nothing like my Dad, which imo is a very good thing.

And no, no one's paying me for my apostasy, lol. I think it's kind of hard when the definition of apostasy within the religion is literally openly disagreeing with anything. Think celebrating a birthday is fine? Boom, you're an apostate. You're literally being influenced by Satan!

While we're on the subject on Hitler, I also hate him. I think part of the reason I had a hard time renoucing my faith was because of this. I heard a ton of stories about that growing up and also about relatives that weren't looked upon kindly here in Canada, either. The "choice" offered to JWs there definitely seems like a deal with the devil, but it's possible to disagree with the JWs and the Nazis. Again, nothing has to be an either/or thing. Nuance is important! Clovermoss (talk) 17:11, 1 July 2022 (UTC)

Yeah, very possible to disagree. I have a friend who I've known for 20+ years and who recently, in his 40s, left the church, when he learned of the details of this. It was interesting because although I respect his freedom of religion, I always thought his was kind of cult-ish, and after 20 years when he finally left and said "they're a cult!" I had this big moment of, "I've been telling you that for 20 years!" Unfortunately in his case, leaving the church also meant leaving his wife, and dealing with excommunication makes noncustodial parenting challenging.

I am so with Karl Marx about organized religion.

I'm still editing Bible though. :-D Levivich^[block] 17:22, 1 July 2022 (UTC)

Eh. Not quite sure I agree with that. Even though I'm an athiest, I think religon can be a force for good, even if it can also have a negative impact. This is a huge part in where I think "reasonable people can disagree" applies. In regards to whole discussion going on at Bible, yeah I agree that people can have such vastly different interpretations and use it to justify practically anything. I agree that there should be more content about how the church basically ruled society for like 1,000 years. Anyways, the bible has inspired people like the Westboro Baptist Church and the Quakers, with such radically different viewpoints. Religon has a potential to harm, but it also has the potential for good. It's complicated. I'd prefer people to just act ethically, regardless of what a book says. My position on this is similar to Hemant Mehta's, but not exactly the same. I will say as someone who had a parent who was disfellowshipped, it definitely was difficult to experience family members shunning them because of what they percieved as love. But it's very possible to disagree, even if I respect someone's choice to believe those things. I'm hesistant to label Jehovah's Witnesses a cult... but I know that some people consider them to be. I can understand that perspective, honestly. I would say that in my opinion, it's a red flag when "how long have you been in The Truth?" is a normal question to ask someone. But where do you draw the line between what's a cult and what isn't? I do care about religious freedom because I think it's an important human right, even though I'm not religious. I'm just super cautious in regards to everything because I don't want to go too much the other way, y'know?

That said, I don't want to drift too much into what Wikipedia is not here. You can always email me if you want to hear more about my philosophical views. Clovermoss (talk) 03:17, 3 July 2022 (UTC)

There are at least a couple of categories that are tricky to address. There's the large silent majority of current users who just edit and don't participate in project space, and there's the population of readers who may become future editors. The WMF has spent effort on the second category in an attempt to expand the editing population, which has caused some experienced editors to decry the lack of attention to base MediaWiki capabilities or in-house support for tools. Organizations can of course work on multiple projects, but to do it sustainably, you have to build up your teams (including project management and long-term support), meaning you have to plan for long-term funding. For a non-profit making money from donations and grants, this means figuring out how to allocate spending between operations now and investing in a long-term trust to generate an ongoing revenue stream. This is all doable (though with overhead expenses), and I'm sure there's plenty to criticize about how the WMF is falling short in its planning. I'm bringing it up just because often users focus on what they think the coding effort will take, and don't consider the total cost of managing a product. isaacl (talk) 13:47, 30 June 2022 (UTC)

Yeah, obviously anything that's at the scale of the WMF has its own unique set of problems. There's also the thing where people focus on what went wrong and not the good. I believe it's called negativity bias? I do think people are trying their best and that their intentions are in the right place. I appreciate you bringing all of this up. Clovermoss (talk) 19:04, 30 June 2022 (UTC)

I do think it's important to address things like WP:THEYCANTHEARYOU though. Clovermoss (talk) 00:07, 1 July 2022 (UTC)

Congratulations on joining the Wall of Reading Text Squad (WORTS)!

Editor recruitment and retention is the whole game. Pardon me for shilling a self-help book but I'm reminded of the thesis of The One Thing: when you have a list of tasks you don't know how to prioritize, ask yourself, "What's the one thing that I can do that will make everything else on this list easier or unnecessary?" For Wikipedia, The One Thing is recruitment. More editors is the one thing we can do that will make all other editing problems easier to deal with or unnecessary: from uncited articles to vandalism patrolling to self-governance to editing user interface... more editors will help all of it, more than almost anything else we can do.

IMO there are two things that are the primary obstacle of editor recruitment: editing atmosphere (incivility), and poor user interface (software). To its credit, the WMF, via UCOC and product development efforts like the Growth Team, is working on both those obstacles, and I think they're doing the best they know how.

One of the problems with the software development -- and it's an example of one of a big general problem with Wikipedia -- is that they don't go about it the way everyone else goes about it; they don't learn the lessons of industry experience. For example, as mentioned above, there is this problem with feedback. The editors who give feedback to the WMF are, indeed, a self-selecting group of interested volunteers, and therefore not representative of the broader community. Also, they don't necessarily "stick" to give full feedback across iterations; instead, you get much more "drive-by hot-takes". It's really hard to get a representative sample of volunteers to give you details feedback across iterations.

Why is it so hard? Because we're not paying them! Duh! In the real world, focus group participants are paid. If you want someone to take hours out of their day to try your product and tell you what they think about it, you need to pay them for their time. The people making the product and conducting the focus group are being paid for their time... why not the participants? Nobody is going to want to spend hours to participate in a focus group for free--and if they do, they're probably a super-user of the product whose opinions won't be representative of the broader user base anyway. The WMF has $400 million dollars, they can afford it; they need to grab a truly random (not self-selecting) sample of active editors and offer them a stipend to participate in a focus group or an extended beta test. That's the only way you'll get meaningful feedback from a truly representative sample. That's why everyone else pays focus group participants and beta testers. (And I don't mean open betas, which is something else.) I know if I have an hour to spend on Wikipedia, I'll spend it on an article or in a discussion with other volunteers before I'll spend it beta testing. Levivich^[block] 16:36, 1 July 2022 (UTC)

Yes, I mentioned focus groups (which I implicitly considered to be paid). It would be tricky getting a representative sample from a community where a large proportion wants to remain anonymous, and many might balk at having to place a paid editor notice on their user page. But all the same it would be worth collecting this info while remaining aware of its limitations. I also think the WMF should partner with universities and take advantage of all that unpaid grad student labour. There is a lot of interesting research into the interactions of online communities or user interface design that could be done which would benefit all parties. isaacl (talk) 20:37, 1 July 2022 (UTC)

Regarding editing environment, to me what is usually considered as civility is a symptom, and not the root issue. English Wikipedia's decision-making traditions don't scale up well, and there's often an advantage to being unco-operative. So that's what we get. Some editors recognize this, but many are content with trying to manage the symptoms in order to keep the advantages they see with the present ways of working. That might be an acceptable trade-off for some, but as you note, it's a limiting factor for growth. isaacl (talk) 21:11, 1 July 2022 (UTC)

Huh. The paid editing thing has interesting implications and is not something that I had even considered. Rereading Wikipedia:Paid-contribution disclosure, I wonder how that would work. I guess people would say their employer is the WMF? Or maybe it'd be more specific like WMF Focus Group? With a link to what exactly that is? My understanding that the main reason paid editing is so tabboo is because you inherently have a COI. But my understanding is that people view stuff like Wikimedian in Residence more positively, but that's something that might not even be directly comparable. Presumably because people are being paid for their expertise and community outreach in a way. Random selection of who gets to be a focus group would be more fair, but I can also see a lot of people being kind of upset about it. Why should they have access to an opportunity that I don't kind of thing? You'd also have the same issues with like sockpuppets and stuff. I'm sure people who think trolling Wikipedia is fun would get a kick out of being randomly selected for something like that and would have no issues trying to make the odds more in their favour, y'know? In regards to the paid contributor thing, since undisclosed paid editing is against the WMF's Terms of Use therotically they could just change that to provide some sort of exception for the focus group. I can kind of see why you would do that, as it could definitely change how an editor's experience is like if other people know they're being paid as a focus group participant. But at the same time, doing anything in a secretive way could seriously backfire and cause controversy. I do kind of get Levivich's earlier point about the people who volunteer their time to provide feedback as it currently is to result in a lot of "drive-by hottakes". Anything an organization does that invites criticism and wants to do the right thing would have to deal with that to some extent, but if you're getting is negative feedback from a small portion of editors who aren't really giving you the feedback you're looking for, I can see like how that can just result in ignoring crucial things (like the whole "talk pages don't matter because most new editors don't use them" example).

In regards to editing environment, I was wondering if we were even on the same track here? I tend to avoid conflict so maybe you're seeing something I don't, but in my experience people who are incivil tend to have anything else that might be relevant disregarded. Because it's not about whose "right". Being incivil becomes the focus of the conversation, which is why civil POV-pushers are something that people care about preventing. Also in general, everyone hates someone whose a jerk. It's just better in every concievable way to be a kind person. So the idea of that somehow there's an advantage to being uncooperative intrigues me. I just don't get how that could be the case, but maybe it's just because the thought has never occurred to me. It's like the antithesis of who I am as a person. Clovermoss (talk) 16:00, 2 July 2022 (UTC)

To clarify, the main reason I care about incivility is that's it's off-putting in a "it adds up" way. Also because it just isn't a nice way to treat your fellow human beings. It leads to people feeling like their contributions aren't valued and what's even the point of contributing to Wikipedia if you can never do anything right? At least that's sort of the impression I've got by observing what new editors who become frustrated and have had a few unpleasant experiences is like. It makes sense, why would they, y'know? It's why it's important to do more than just not "bite" the newbies. But at the same time, maybe that's the advantage to being uncooperative, I guess? I don't know. It just seems like a loss to everyone involved. But if someone really cares about something in particular, in a twisted way I can kind of see that being an advantage? As in, they won because there's no one left to disagree with them? But that's just... horrifying. I'd like to think most people who edit here don't think that way. Clovermoss (talk) 16:17, 2 July 2022 (UTC)

I've considered writing an essay "Please don't be obnoxious, even if there's no rule against it," but it would be unlikely to reach those who could benefit from the advice (at least not in an effective way). Instad, I wrote a few sentences for a possible essay called "On kindness" with the thesis statement, "If you're planning to be part of the Wikipedia community, exhibiting kindness towards others has better long-term benefits than un-collaborative behaviours." The problem is that the long term... might be very long. The key issue is that most editors aren't interested in conflict—like you say, you avoid it. If someone is being unpleasant, this is just a web site, so the natural tendency is to go somewhere else. Decisions are made by a self-selected sampling of the community that happen to show up in the right place at the right time. So there's incentive for escalating aggression in order to drive others away, or to induce them to do something overly aggressive, and thus have an editing restriction imposed. The "consensus can change" principle means editors can keep doing this as long as they have patience for it. The good will of collaborative editors gets taken advantage of, time and time again. isaacl (talk) 20:48, 2 July 2022 (UTC)

Yeah, I get what you're saying. I think that the problem is that inherently, the people who are 100% convinced in their position on anything are going to be the people who are the most willing to do whatever it takes for that to happen. I know I used to be very good at that when I was a JW because I literally felt like nothing could stop me, not even Satan. And Satan controlled this system of things, so... My point is that having doubts about anything makes it harder to be that adament about something. Because you're trying to think and consider other perspectives, which is a good thing. It's why stuff like gish galloping is effective. It takes time to counteract things and it's exhausting to do that point by point. What's the point of doing so if you're not even a politician or convinced that you're somehow saving people from imminent destruction? I don't want to make comparisons to my religious upbringing too much, but sometimes it feels relevant, and as I said, it is part of what made me the person I am today. That said, this isn't really directly comparable. I hate generalizing and care about nuance. It's a comparison, but it isn't really quite the same. It's just hard to describe how to feels to be 100% certain about something and I'm not sure anyone here really advocates for things that way. I guess my point is that it can be like that, though not quite like that, sometimes? Maybe? I don't know.

Anyways, strong insistence can come from a place with good intentions but it's really off-putting to whoever has a different position but isn't willing to argue about it indefinitely because they just don't care about it that much y'know? People have lives and typically care about things beyond Wikipedia. So you're inherently going to end up with a bunch of decisions where people who care enough to keep the argument going eventually win. I can see this resulting in whoever participates casually but has an opinion basically being ignored in most circumstances. I agree that kindness is the best long-term. And also, just in general. But I also think the premise of some editor retention stuff is also kind of unrealistic? Very few people are going to care about Wikipedia enough to make 100+ edits a month for several years of their lives. It's a lot to ask of someone, even if there's the rare person who thinks it's fun and does it for whatever reason. We really do need more people who do contribute casually. I don't remember the exact calculations of what I did at the village pump discussion about my out-of-the-box idea, but I vaguely remember something along the lines of if 0.1% of readers added a citation to an unsourced statement, we'd basically be rid of the huge overwhelming backlog in two days. But if I wanted to do that entirely by myself, I'd have to fix 10 every day for 109 years. Clovermoss (talk) 00:08, 3 July 2022 (UTC)

I'm not sure by "editor retention stuff" if you are referring to how it's generally discussed on, say, the editor retention WikiProject talk page, or metrics from the WMF. I think, though, that no one thinks editors should be roped into contributing for longer than they want to. The amount of content can be increased through increasing the top line and the bottom line: recruit more editors, and make the collaborative process sufficiently attractive so editors remain interested in contributing. There are, though, limits: not everyone is well-suited for writing or evaluating encyclopedia articles for a general audience with a neutral point of view. And writing in a group is just hard—in the real world, generally the work gets broken down into separate parts that are delegated to individuals and then reviewed together after. Trying to work through issues through talk page discussions is extremely challenging where not everyone may be available at the same time, individual editors can, even in good faith, swamp discussion, and often editors lack patience to let everyone have their say and work towards agreement. isaacl (talk) 22:08, 3 July 2022 (UTC)

I was talking about it more in a general way, I know that the editor retention WikiProject talk page tends to more realistic. But I've kind of got the impression that people care more about what ~4,500 people making 100+ edits a month are doing and not the 36,000 people who make more than 5. [1] The latter aren't really considered as "active editors" and 100+ edits seems to be the threshold where people kind of care about your opinion. But also tenure? My general point is that less than 5,000 people being involved in what makes Wikipedia function as a website that literally millions of people read isn't that sustainable and I kind of get why people are concerned about that. But at the same time, asking people to devote even more time (potentially even years of their lives) to Wikipedia isn't always the best option? People have other things to do in their lives and very few people are going to be interested in making this website their main hobby. According to this [2] I've literally spent more than 700 hours here and I still don't quite yet understand everything that goes on. Smart Serve, the local alcohol serving certification in my jurisidiction, can be completed in about 4.

My overall attitude hasn't changed much since I started editing and I've always looked before I leapt but I feel like people take me more seriously now simply because I've been around awhile. I was also referring to the general rhetoric of "the sky is falling, we're losing our active editors"... which, to some result, is true (but metrics like Wikipedia:Time Between Edits seem to suggest that the situation may not be as bad as it appears to be). Backlog drives tend to be almost a bandage on overall issues, I guess? Ideally they wouldn't get to the point where there are massive backlogs because there'd be enough people working on them to make that not an issue. This can also lead to people doing so much that they become burned out. Wikipedia by its very nature is going to lead to people focusing on whatever their specific interests are by its very nature as y'know, a place where people volunteer their time. Personally, I deal with enough conflict in my real life job. People don't always treat young people working minimum wage jobs that well. A lot of the time I don't feel the need to go much beyond my own personal interests, although occassionally I do. Part of that is the fear I'm going to miss some crucial detail or mess things up because I'm 19 and the older I get, the more I realize I just don't know about the world. Wikipedia has a lot of potential real-world implications, too.

I think that's part of the reason there's always been an issue with NPP/copyright violations/AfD or whatever being backlogged is because people are way less likely to find that a valuable use of their free time. We shouldn't expect people to have to. I really doubt I'm the first person to point this out. People talk about how we need to reduce x or y backlog, and yes, that is something that should be done, but at the same time I agree with you and Levivich is that the answer to that is more editors. More editors basically solves all of these issues to some extent, so that should be the priority. But Wikipedia editing in general has a high barrier to entry. I'm a fast reader and I know how to use the "find" feature of my browser, but most people are not interested in reading stuff like the Manual of Style for fun. Let alone the absurdly long Village Pump/RfC discussions that go on. I think Joe Roe made an interesting point in the growth team discussion about how stuff like VisualEditor can help people focus more on content instead of trying to mess around with formatting. In the grand scheme of things, very few readers care about things like portals, categories, short descriptions, etc. Not that those contributions aren't valuable, because they are, but it's not what the average person imagines when they think about what Wikipedia is all about. Ideally people shouldn't have to spend time being confused about how our weird markup language works before writing about whatever obscure topic they're passionate about. Clovermoss (talk) 23:19, 3 July 2022 (UTC)

My guess is that editors who make small numbers of edits annually are likely making typographical and grammatical fixes, and introducing small fact updates or corrections. (You see something wrong or perhaps missing, and you fix it.) This is indeed highly valuable in the aggregate, because it forms the long tail in the distribution of editors. Other important tasks, like new page patrolling and copyright investigations, are more likely to be performed by more vested editors, and so they'll naturally have higher edit counts. Attention should be paid to both groups, particularly since the first group is a potential source of editors for the second.

Although there is a lot of written guidance available, if a potential new editor understands what it means to write an encyclopedia article and how all info should be traceable to sources, they don't really need to read any of it ahead of time. When a writing style question crops up, they just need to know how to look up an answer. (Knowing about WikiProjects would in theory help them find other editors to collaborate with, though sadly many WikiProject talk pages no longer get a lot of responses.) It's why I think the success rate at keeping an editor around for some period of time can be improved if we try to recruit from communities that do know what it means to write an unbiased and sourced record of history. isaacl (talk) 00:55, 4 July 2022 (UTC)

You stated what I was trying to explain a lot better than I could. People are capable of teaching themselves what Wikipedia is about but a lot of this information isn't easy to navigate. If you figure out how to do that, you have a good foundation. But there's a million little things that can kind of get in your way at the same time? If you don't edit somewhere more controversial, it's like your edits go into a void unless you actively seek input from others (which is what I did). If you don't do anything problematic you could basically go unnoticed for years if you fix a typo here and there occasionally, even if you may be interested in contributing more if you knew how. I had a similar experience with TVtropes, actually, now that I'm thinking of it. I was able to fix typos because that's something that was easy enough to fix but I never really did much beyond that because I couldn't figure out how CamelCase and everything else in their editing interface worked. You might find the previous discussion about new editor experiences that Levivich, HLHJ and I had here interesting [3]. I had an experience that was within what is suggested at welcome the newcomers. My understanding is that Levivich's wasn't as welcoming based off the comments there, although maybe they reflect on their experiences differently now? I can't speak for other people. My first edit was literally fixing a typo, but I was also really insistent on asking other people for advice right from the start. I'd say the main benefit my previous wikiHow experience had (apart from some familarity with MediaWiki format, even if it wasn't really enough to not get confused about many things here) is that I was aware that Wikipedia also likely had a community "behind the scenes". This meant I could ask for help before I did anything I was unsure of and I didn't really ruffle any feathers. A lack of feedback isn't nessecarily a good thing or a bad thing, but it can be disheartening and it can also lead to problems down the road. Maybe it's just me, but I'm still working on my confidence in that I actually know what I'm doing here and trying to alleviate some of the gaps on what I know I'm not that good at. I think part of this has to do with having really high expectations for myself in general, but at the same time even though I've spent a lot of time lurking/reading what other people say and trying to figure out what works here, actually doing things based on that judgement can get a bit more complicated. Clovermoss (talk) 03:17, 4 July 2022 (UTC)

It's interesting to hear about other editors' experiences. I had no expectation of getting any feedback so it wasn't an issue for me. I can't remember how I learned about discussion on the talk page; that is one important thing for new editors to know. Upon further reflection, I can see a good bit of luck affecting an editor's first impressions. If you happen to edit an article being watched by someone who is protective of it, you might find yourself getting reverted with a brusque, dismissive comment. Even if somewhat warranted, it can be very off-putting (and if not warranted, it feels like a waste of time). I hardly edit main space now for a variety of reasons, but not least of which is that the potential for conflict over even the most benign edits is wearying. isaacl (talk) 04:40, 4 July 2022 (UTC)

I think the other thing is that people usually are focusing on preventing harm and thus are at best apathetic to the vast majority of new editors. I have an unconvential approach of how I sometimes patrol recent changes (since you can filter it to show likely good faith edits by newcomers). The main reason I do this is because I always figured it's much better to try to approach someone while they're still in the middle of the first (or maybe only) attempt at editing and there's already a huge focus on people who are keeping an eye on vandalism so it seemed like something I could have more of an impact doing, y'know? I also try to welcome people I see making good faith edits on my watchlist. Clovermoss (talk) 05:13, 4 July 2022 (UTC)

In regards to talk pages in general though, I was actually referring to what Levivich said in the growth team discussion: It wasn't that long ago that we learned that the reason IP editors didn't see links to article talk pages is because the WMF thought that it wasn't important for IP editors to use talk pages. (!) They did a study and found that using or not using talk pages doesn't meaningfully change the frequency that an editor is reverted, and that's how they decided talk pages weren't important. These decision, and others like it, were made by people who do not edit Wikipedia. That's why wiki-apps written by editors are better than what the WMF puts out. [4]

Anyways, I should also probably ping the only person (at least that I'm aware of) from the WMF that's posted on my talk page: MMiller (WMF). They might have something interesting to say, since we haven't really asked anyone from the WMF what their thoughts are. They're also involved in the actual growth team stuff so they might have some interesting thoughts about that, too. Clovermoss (talk) 20:15, 4 July 2022 (UTC)

I wasn't referring to anything you said when I mentioned talk pages; just commenting on something I think new editors have to learn. I wasn't intending to start a multi-person conversation on general thoughts regarding editor recruitment and retention, so I'm not sure how much further I'll go down this digression. isaacl (talk) 23:25, 4 July 2022 (UTC)

Oh, okay. That's fine. I thought you were mentioning them as a sort of reply to one of my previous comments but the discussion kind of went in a million different directions after that, so I thought I'd bring that up in case it was relevant? I'm not saying that I nessecarily agree with all of that, just that there should be better communication between the WMF/the general community and that's an example of where things didn't really go that well. Regardless of whether or not you continue this discussion (I think it's probably going to cease sometime soon because they're only so much discuss on this topic because it's something that people have had a lot of discussions about in general)... I did want to say that I really appreciated everything you said and your perspective on things. You've clearly put a lot of thought into everything. Clovermoss (talk) 00:17, 5 July 2022 (UTC)

Thanks for the ping, @Clovermoss. I am definitely interested in conversations and thoughts about new editors! I'm about to be away from work for a week or so, and I'd like to be able to read this very long conversation before saying anything in particular, so I hope to get back to you a bit later. I hope that's okay! MMiller (WMF) (talk) 15:17, 7 July 2022 (UTC)

@MMiller (WMF): That's completely fine. It is a gigantic wall of text. There's also a link to a previous discussion at the top that you might be interested in. Enjoy your time off work! Clovermoss (talk) 15:56, 7 July 2022 (UTC)

There's also this discussion where I wrote a bit more about what exactly I meant by the 1Person1Ref idea. Specifically in the third subsection about out of the box ideas. Clovermoss (talk) 16:14, 7 July 2022 (UTC)

@MMiller (WMF): Are you still interested in having anything to say? Or are you still going through it all? If it helps, I don't expect you to say anything about the religious stuff that just kind of poured out of me. I guess I've been thinking more about my life experiences and how that's impacted who I am as a person more lately. In regards to everything else, I truly am willing to listen to your perspective even if it's quite different. I've always admired what you've done specifically, and maybe part of thw problem is that it's relatively easy to conflate the WMF as a whole with anyone whose willing to listen, and that's not really that fair to you. Clovermoss (talk) 10:16, 18 July 2022 (UTC)

Hello @Clovermoss. I've read over these interesting conversations you started! Thank you for being enthusiastic and thinking hard about the health of the wikis. I think it's definitely reasonable to be asking about how the product teams at WMF might be more productive, and you might be surprised to hear that I am always asking myself that same question (it's part of my job!) I definitely feel a lot of urgency to get major things done quickly for our users -- the internet is changing fast, and we want free knowledge to be there for people in five or ten years time and beyond. As this conversation has touched on, though, shipping changes to our software turns out to be quite complicated. For instance, whenever we build something new or change something old, we need to consider how it will work in multiple skins (e.g. Vector, MonoBook, etc), how it will work in multiple languages (e.g. English, Arabic, Korean), and whether/how it will be usable on both desktop and mobile. I definitely carry what I would consider a "healthy impatience" to make an impact. I know we could have a much longer conversation about it all!

I'm glad you feel like the Growth features for newcomers are on the right track. We've done a lot of research on what makes newcomers stick around and what makes them leave, and you touched on a couple of the big ones. Civility (or its opposite, toxicity) definitely matters. You might be interested to read the New Editor Experiences research, which interviewed dozens of newcomers and confirmed that many newcomers are turned off from Wikipedia after doing something wrong and then getting made to feel like they're not welcome. And you also touched on how we need to bring many more people into editing. One of the main ways the Growth team has been trying to do that over the last couple years is by creating new ways to edit. To edit successfully, newcomers need to learn wikitext (or the Visual Editor) along with so many policies and rules. Many people who might be interested in contributing to Wikipedia don't have the time or technical skills or tenacity to do all that learning and trial-and-error. But we still want to give them the opportunity to try editing and perhaps get the spark of enthusiasm that prompts them to go deeper. That's why the team has developed two "structured tasks", which are "add a link" and "add an image". They are streamlined editing workflows that newcomers can do without too much skill or learning required. They're not yet on English Wikipedia, and I would definitely be interested to know what you think!

On that note, I wanted to mention how we do research and learning. You talked about things like focus groups and bias around the information that goes into the product process. Yes, we totally do many different kinds of research as we plan our work! This includes interviews, live user tests, surveys, office hours, and community conversations. That New Editor Experiences research that I linked above is an example of a time that WMF spent money on recruiting interviewees, translators, and researchers to try to create an unbiased picture of newcomer experiences. Another example is these prototype tests done by the Web team around the Desktop Improvements project. One of the reasons we gather information through multiple channels simultaneously is to help mitigate biases that could come from any one of them. For instance, when thinking about Growth features for newcomers, it obviously wouldn't make sense to only test the features amongst experienced editors (even though those are the most readily accessible users), because they're not the target audience. So we seek out newcomers for testing. But we do also test the features with experienced editors, because they'll be able to see things from their perspectives that newcomers can't -- perhaps problems the features could cause further down the line for patrollers.

Well, I wanted to touch on a few of the major points you brought up, Clovermoss. I hope this is interesting and helps! MMiller (WMF) (talk) 23:59, 20 July 2022 (UTC)

There's another example of this phenomenon playing out now; I recently added to WP:CENT a link to the village pump proposal about showing edit notices to mobile editors. This is another thing -- mobile editors don't see edit notices -- that's been flagged by the community forever (there's a phab ticket from I don't remember how many years ago), and the WMF just thought it was low priority. In the meantime, a volunteer editor wrote a hack that shows edit notices to mobile editors, and the community is about to implement said hack. Once again, the community just steps in to fill an obvious gap because the WMF just didn't think it was a priority. The WMF has $400 million dollars, but volunteers on the internet were able to fix this faster. It's evidence of the poor prioritizing of WMF dollars.

Another example: I don't participate in the m:Community Wishlist Survey because I think it's BS, given how the prioritization is decided by the WMF. We can see at m:Community Wishlist Survey/Updates/2022 results#The leaderboard that the community's votes constitute the "popularity rank", which is one of four criteria. The other three criteria are set by the WMF. Then the WMF calculates a final "prioritization score". Now, it's fine for the WMF to weigh the community's priorities against other criteria, such as feasibility, complexity, etc. But one of those three WMF-set criteria is "community impact score". What the heck is the difference between "community impact" and "popularity"? Simple: the WMF wants to decide the community impact, rather than just letting the community decide that by voting on things -- this, the WMF calls popularity. Each item's final score, the "Prioritization score", is calculated by some algorithm that does not appear to be disclosed by the WMF (at least I can't find it). The end result? Click on the leaderboard link and sort the table by "popularity rank" or by "prioritization score" and you'll see there is basically no correlation at all. In other words, our votes don't matter, they're actually wiped away via some calculation determined by the WMF. So I'm not going to spend over an hour reading and thinking about the Wishlist Survey when I know the WMF will just set the priorities based on what they think is the "community impact".

This is why people say the WMF doesn't listen to the community. Based on interactions with WMF devs, I, like others, have come to the conclusion that the devs do not think the community knows what's best; they think they do, and that's what drives software development priorities. That's why we have Wikimedia Enterprise (which BTW is losing money and behind schedule) before we have a decent WYSIWYG editor. Levivich^[block] 18:15, 7 July 2022 (UTC)

I did want to mention that I don't think they're completely unaware of the tension that can exist between the WMF and the community. They also have a volunteer account that they use to edit, Cloud atlas.

I can't speak for everyone who edits here and I doubt they can speak for the entirety of the WMF. I do think broader awareness of these issues is important though, with more communication and awareness being vital for everyone involved. Clovermoss (talk) 19:17, 7 July 2022 (UTC)

Yeah, true. Cloud atlas has started multiple new articles, and brought multiple articles to GA (thank you!). That's an imperfect metric, but I wonder whether that's typical or atypical for WMF staff. I know a lot of WMF staff have personal accounts and edit, but I wonder how many of them have GAs, or similar, like how many have done "heavy duty content work". Levivich^[block] 19:36, 7 July 2022 (UTC)

As I posted in that discussion, I think research is great and something that the WMF should do, but shouldn't hold up immediate needs. Prioritization challenges, though, are not the sole provenance of the MediaWiki development team. The interface admins, for example, resist implementing per-page gadget loading in Javascript, as they disagree with the principle of having the client take on the workload of making this decision. (Maybe they're right to defend this design choice against popular opinion from editors that the performance cost is minimal, maybe they're not.) So as much as I do think a better prioritization mechanism should be put into place by the WMF, with more funds allocated towards maintenance and unglamourous fix-up projects (I do know it's not as easy to do as it is to say), I acknowledge there is a role for code owners (in the software development sense) to act as guardians of key design decisions. Of course, there should be occasions when those design decisions are re-examined to see if they are still applicable and desirable. isaacl (talk) 20:39, 7 July 2022 (UTC)

As for Wikipedia Enterprise, I don't think that's an example of the development team thinking they know what's best. I think that's an example of how it's easier to decide how much investment you want to spend on a product that is forecast to have a revenue stream, than one that has no revenue stream. That decision is made at a strategic level, above the development team. isaacl (talk) 20:42, 7 July 2022 (UTC)

Arbitrary break since this conversation is huge and still ongoing

@MMiller (WMF): Thanks for getting back to me. I'm trying to take it easy on myself due to stressful real life events, but I do have some thoughts about what you just said. So... I think a generalization I've seen is that new editors experiences are either good, really bad, or just kind of apathetic. If you're doing something wrong, you're definitely going to get feedback because people are focused on minimizing harm or just keeping with established consensus that you might not be aware of because there's been however many discussions over the past two decades. I wasn't aware of MOS:CELESTIALBODIES until today when I was like "it's kind of weird that Sun is capitalized sometimes" and my change was undone. I didn't know that and it was a good faith mistake, but you're way more likely to have something like that happen if you don't even know that the MOS is a thing or how to reference it.

There's also the whole "Wikipedia abbreviations" thing. We basically have a dedicated subculture where things can be a bit different than their standard English meanings or even inherently contradictory. A lot of the essay pages are like that too. They have a disclaimer that they can represent widespread or minitory viewpoints, but it's difficult to tell when you're reading them just where anything might fall on that. Then there's individual discretion on where to draw the lines. There's a million links to everything. I think something I have agreed with in regards to Sdkb is that streamlining welcome templates is a good idea because it gives people some sort of basis to be like "this is what you do" that doesn't contradict itself or give you massive walls of text that prompt you to TLDR. I think there's also just some general issues with banner blindness too. A lot of stuff about the Wikipedia community and editing and everything (especially on the main page) is hidden by things that you basically instinctually ignore if you grew up in the world where the Internet was a thing. You pay the same amount of attention to them as you do to pop-up ads... something that's just not worth your time and doesn't even register. Like I'm looking at the bottom of this very page and I'm like "oh there's this toolbar with characters like → I just straight up ignored for years". There's also a lot of issues with the mobile version of the website. I'm like Cullen328 in the "when I edit on my phone (like right now) I use desktop view for basic functionality" way.

Going back to my good efforts disappear into a void comment earlier on in the thread, the impact feature that's in the growth newsletter seems like an interesting step in the right direction of doing something to help mitigate that. As for wikilinks and images... it might help with some familiarity with wikitext, but there's still a lot of uphill climbs from there. There's also stuff like MOS:REPEATLINK that might have the outcome of newcomer tasks like that kind of backfire. Unless I'm misunderstanding how that task works. I think that while our attention on retaining new editors needs to be improved we also need to give more attention to intermediate editors too, since imo they're the most likely to stick around. Not everyone is interested in making editing Wikipedia their main hobby and that's okay. I do think trying to create some sort of stepping stone with newcomer tasks like that in general is a good idea, though. Clovermoss (talk) 01:05, 21 July 2022 (UTC)

I am commenting because I was pinged. For over ten years, I have done over 99% of my editing on Android smartphones and I almost always use the fully functional (and improperly named) "desktop" site on my phone. This is a collaborative project. These mobile sites and apps have been in development for many years and they are still an active impediment to collaborative editing. So, countless millions of dollars have been squandered for well over a decade on failed ventures, all while the desktop site works perfectly fine on modern smartphones. It is Kafkaesque. Cullen328 (talk) 02:03, 21 July 2022 (UTC)

Pinging MMiller (WMF). Just so you do not think that I am somehow kind of ignorant rando, I want you to know that I have written roughly 100 articles on my smartphone, none of which has been deleted. I have substantially expanded many hundreds of articles, and just today, dramatically expanded and referenced Clambake, all on my phone. I truly love improving encyclopedia articles on my phone. I have taken articles to Good article status, on my phone. I am among the most active editors at the Teahouse and the Help desk, on my phone. I agreed to put my name forward for administrator on my phone and was approved by the community by a vote of 316/2/3. As an administrator, I have blocked 6402 vandals, trolls, spammers and copyright violators, all on my phone. I protect countless articles from vandalism, all on my phone. Almost every single day, I devote serious volunteer time, all on my commonplace Android smartphone.

Over and over again, for year after year, I have offered to discuss my smartphone editing expertise with WMF staff. I lived 35 miles from WMF headquarters for many years. Crickets. I now live more like 135 miles from San Francisco. Crickets. I now live in the same small town as a WMF board member. Crickets.

Why does the WMF entirely ignore the input of highly experienced and highly productive volunteer editors who have made objectively substantive improvements to this encyclopedia, and have detailed insights into the specific problems? My hypothesis is that "friends might lose their jobs" but I am fully prepared to be convinced that my assessment is wrong in this case. Cullen328 (talk) 02:32, 21 July 2022 (UTC)

Hello @Cullen328 -- thank you for being such a dedicated editor, especially from your mobile device (and for helping tons of newcomers in the Teahouse!) I just read your essay about how you edit, and I would definitely like to discuss more. I'm sure we have things to learn from your experience and could generate ideas to improve mobile editing. I'll send you an email to see how we can connect. MMiller (WMF) (talk) 00:45, 30 July 2022 (UTC)

Accessing the desktop site on mobile doesn't work all that well. It's hard to use. You have to scroll alot and zoom to be able to click on things. It's not optimized for mobile design. I would always edit using the mobile version, flawed as it is, unless I needed some feature that isn't there (the fact that some aren't is bad design). The messages thing needs to be fixed though. (t · c) buidhe 02:32, 21 July 2022 (UTC)

Buidhe, I deeply disagree that it doesn't work all that well. Scolling and zooming are commonplace behaviors on smartphones, and using the desktop site on mobile rapidly becomes very easy for those who try. Plus, the desktop site is fully functional on Android smartphones. It works right. Some people say "well, you must have great vision". False. I am 70 years old. I have amblyopia that left one of my eyes almost blind. I have cataracts. I had cataract surgery on my good eye because operating on my bad eye would have been a waste of time and money. I have low tension glaucoma and have had a laser procedure to help with that. Despite all of that, I have no problem at all editing on my Android smartphone, as I am doing at this very moment. Cullen328 (talk) 02:48, 21 July 2022 (UTC)

My near vision is fine and I have trouble seeing all the text let alone clicking on it without a lot of zooming and scrolling side to side. I do edit this way a fair amount, mainly because the mobile site doesn't support citation template generation, but it is considerably more difficult for me. Admittedly my smartphone is on the smaller side. I'm glad it works for you, but I don't think that all or most mobile users would be able to access Wikipedia easily using this method. (t · c) buidhe 02:55, 21 July 2022 (UTC)

I also prefer the desktop view but I have edited using the mobile website (and even the actual app which was not ideal for anything beyond basic typo fixes at least in 2019 [5]). I also don't mind the scrolling but I'm used to it. I actually dislike the automatic hidden sections in mobile view, but maybe that's just me. As for eyesight, I don't wear glasses and I'm young so maybe I'd have a different opinion if I wasn't like that. But I also get annoying popups from my browser (Chrome) that try to "simplify" the page when I use mobile view. Maybe there's something I could do in settings somewhere to fix that. But I will say that I haven't heard often from people who say that they actually like the mobile site and that's an interesting perspective. Ancedotally my experience has been that people agree with Cullen328 most of the time. It might be worthwhile to actually try and survey people in Category:Wikipedians who edit by smartphone for their opinions, though. Clovermoss (talk) 03:50, 21 July 2022 (UTC)

@Buidhe: so you inspired me to experiment with mobile view a bit more. Haven't tried to do anything outside of mainspace yet, but I was able to manually add a citeweb template. [6]

I don't think a new editor could easily do that because you basically have to have the parameters memorized for it to work... but still, not impossible. The overall experience seems to be a bit better than what I remember though. Maybe I was mixing up my impressions of the app with mobile view in general? Or the experience has been improved since I last tried? It's still not my favourite way to edit, but my brief experimentation has lead to me belive it's tolerable. Maybe I could get used to it. Clovermoss (talk) 06:35, 22 July 2022 (UTC)

for a new user it has most of the necessary functionality (they are more likely trying to correct a typo than add a citation) with added convenience of bigger buttons, larger text, lack of scrolling/zooming etc. I don't support de-defaulting the mobile site although users should have the option to switch back and forth. (t · c) buidhe 06:38, 22 July 2022 (UTC)

@Buidhe: Agreed on that. That's actually mostly how my edits with mobile view/the app have worked in the past. Something small enough to fix that seemed like it wouldn't to be too difficult to accomplish with that interface. I was more just seeing how a more "serious" editing experience might look like. I might even try downloading the app again... I'm curious about the differences, honestly. I still think the desktop view is ideal but I wouldn't say that everything else is completely broken (at least based off my initial impressions).

Agree with Levivich in one of the prior conversations that the WMF has a lot of resources at its potential and could have used them more effectively... but we got to work with what's happened already and go from there, I guess. I really do have some faith in MMiller (WMF) and I recognize that he can't speak for the entirety of the WMF and that his efforts mostly focus on new editor engagement specifically. I think part of the problem in general is a lack of suistained communication with the greater community and a lack of trust from previous well-known issues that is hard to regain. That leads to people feeling frustrated and that the status quo will never change. I haven't personally been greatly disappointed yet, so maybe I'll feel differently in the future. So far I like having hope, though. I'm also kind of out of touch with a lot of things myself so maybe I'd feel differently if I was more experienced in certain areas. I have a tendency to feel like I half-understand things even though I'm trying to work on that. I don't really contribute to policy-related discussions much. Clovermoss (talk) 06:53, 22 July 2022 (UTC)

Leaving this here in case anyone is interested: User:Clovermoss/Mobile editing. It's still very much a work in progress. Clovermoss (talk) 00:50, 24 July 2022 (UTC)

Hi @Clovermoss and everyone in this conversation -- thanks for continuing to talk about newcomers and mobile editing. As we can all tell, these are such big challenges. I've learned a lot about it all over my last four years at WMF, being exposed to lots of data and user research studies. Here are a couple things on my mind as we talk about this:

• One of our biggest challenges is that we're trying to build software for every kind of person around the world. Since this is meant to be the "encyclopedia that anyone can edit", we find ourselves considering how people of all different ages, locations, education levels, and with different types of devices need to edit. We learn about users in Africa, in India, in Southeast Asia, in Latin America, in the Middle East, and we discover that they have varying expectations of what a good mobile experience needs to be for them. It definitely makes it tough to settle on designs that are a great fit for lots of people, but we're always trying.
• When it comes to mobile, I sometimes think about how people these days do all sorts of complicated and consequential things from mobile devices: they pay their taxes, apply for mortgages, buy cars, etc. And so a natural question is why more people don't edit Wikipedia from mobile. Something that comes to my mind is that in the cases of taxes, mortgages, and cars, the interfaces lay out the process as a series of discrete steps. If you're doing your taxes, the software might say, "Enter your income", then you push "Continue", then it might ask, "What state did you work in?" -- all one step at a time. But when it comes to editing Wikipedia, there isn't a clear "one step at a time" sort of path. There are an infinite number of ways you can change a Wikipedia article, and our software (visual editor and wikitext editor) let you do all sorts of things in all sorts of orders. We've learned that that is an overwhelming experience for many newcomers, and may be especially tough on the small mobile screen, because on mobile, people are used to the "one step at a time" sort of workflow. That's the origin for the idea of the "structured tasks" I mentioned above: we're attempting to break down editing activities in a series of steps that are easier to do on a mobile device and that guide a newcomer through the process. Does this make sense?

MMiller (WMF) (talk) 00:38, 30 July 2022 (UTC)

@MMiller (WMF) Yeah, I get what you're saying. I've been experimenting with the app more lately and it's been a bit of a bizzare experience because there isn't really clear instructions. I'm used to editing Wikipedia at this point, but not like this. Do you have any idea what the "perfect edit quality" thing is, for example? My best guess is that it's based off some percentage of "live" or non-reverted edits, but maybe I'm wrong. I've had some very brief impressions so far, some good and some bad. I'm trying to keep an open mind even if a lot of stuff doesn't seem ideal from my perspective (and I think that's a bit of understatement on my part, I've heard lots of terrible things about the app in general from pretty much every experienced editor, even the ones who are open to mobile editing). Wikipedia editing itself isn't exactly straightforward but the app feels even more out of place in a way? It's difficult to access things like the wikipedia namespace if you don't know how to search for the exact page you're looking for. There's no obvious instruction pages. Small tasks are something and I think that overall they're a good something, but a lot of that doesn't nessecarily seem super clear either. I was very confused when I was prompted to an "article description" to a usersubpage, that "quick facts" were collapsible infoboxes, being prompted to add a caption to an image that already had one, etc. I do get that there's a lot you have to consider though and I appreciate that. I really do think more open/enagaged conversations like this would help with the rift that people can feel in regards to the way the community and the WMF in general interact. Thank you for your comments here, I appreciate them. Clovermoss (talk) 01:44, 30 July 2022 (UTC)

Just to clarify, @Clovermoss, it sounds like you're using the Android app and its suggested edits, is that right? When I've been talking about mobile editing, I've mostly been thinking about the mobile website that you would get in the web browser. In the Android app, yes, it is using revert rate to communicate about your edit quality. That's meant to help newcomers understand that quality matters and is judged by fellow community members. MMiller (WMF) (talk) 01:57, 30 July 2022 (UTC)

@MMiller (WMF) I've mostly been experimenting with the android app lately, but I have some experience with the mobile version of the website, too. Almost all of my mobile editing that I've done the past few years has been through desktop view regardless of what device I've been using. This conversation inspired me to try new things and identify what stood out to me as a more experienced editor. There's more at User:Clovermoss/Mobile editing, but again, still in the very early stages of whatever it is I'm doing. I started all of this like maybe a week ago at best? I'd have to look through my contributions for a more exact date. My point was more along the lines of nothing explains what the perfect edit quality is exactly and people might react to that information in unexpected ways. Clovermoss (talk) 02:44, 30 July 2022 (UTC)

Wow this is a varied discussion! Attempt to contribute:

1. There is a massive waste of Cullen328's time. Way back, like over a decade ago, you could paste a URL between single square brackets, hit "Save", and a bot would later turn up to turn your bare URL into an autofilled citation template. For academic articles, recently-published books, and other sources with computer-readable structured metadata, it worked perfectly. For more casual websites, it usually worked pretty well. Then the bot died somehow, and later WP:Reftoolbar came instead. This was inferior to the bot as it required me to sit and supervise it while it autofilled the ref (and sometimes it takes many seconds), and because I had to hit a special button to raise its interface. I think User:Mvolz, who has worked on citation tools, said it was based on the Zotero scrapers. Our current Help:Citation tools are varied, but given the choice I'd opt-in to a bot that just expanded my URLs after X min of inactivity on an article (so it didn't edit-conflict me). I think User:Cullen328 might too; it would certainly make editing faster for him. Improving the efficiency of editing is a good way of increasing available editing time,
2. I am not everyone, but I hate step-by-step interfaces of the sort MMiller (WMF) describes. In my experience, they are generally used to coerce me. They hide the later stages of a process from me until I have sunk substantial time into the early stages. They almost always force me to complete each stage, no matter how irrelevant and timewasting. They usually don't actually allow me to do what I want to do; they only work for very very standard use cases. They are the web version of telephone menus: "Did you know you can find lots of useful information on our website? We are currently experiencing unusually high call volumes. For our hours, press one...". If I wanted the hours, I would have checked the website! A huge thing I love about editing here is creative freedom, the freedom to choose my own tasks and how to do them. If I had to edit through a step-by-step interface I wouldn't edit. If it had been the first option presented to me as a new editor, I would never have become a regular editor. Buying a car is not creative work; it is innately highly-structured. Nothing against editors who do highly-structured tasks, thry are highly necessary, but it's not for everyone.
3. I tried to edit using Firefox on a Pinephone (proper mainline Linux, not Android, so the app might not work). Mobile editing was painfully slow. I could do almost everything, but my productivity plummeted. It didn't look like practice would fix, either. I think mobile interfaces are going to improve, at the mobile end, tho. When I next edit from mobile, I'll try using the desktop page, thanks for the tip.
4. Clovermoss highlights an interesting issue; acculturation. If a tool for new editors does not help people learn the norms of the community, if it teaches them jargon and concepts and attitudes that confuse established editors, that is not good.
5. Tools that create permutational complexity make everything harder to maintain. Modularity simplifies. Volunteers typically make modular tools. Markup helps with that.
6. I don't manage to contribute to all the top-level discussions I really care about. Life happens. I don't even hear about all of them. Result, I feel disenfranchised and chivvied.
7. Being rude and nasty and bullying and uncivil can be used to manipulate people; it means that opposing the jerk is really unpleasant. Making the negativity of the response unpredictable decreases the amount of effort-put-into-being-nasty needed for the same deterrent effect. I can dig up sources on this if wanted. The first really nasty feedback I recall on WP was from an editor who later turned out to be a paid shill and got banned; they seemed to consciously be using unpleasantness to discourage contributors. My nastiest interactions have been on high-stakes medicine- and specifically nicotine-related topics. But a lot of hostility is known issues with text-based communications; the community's neurodiversity also creates an acculturation issue for some.
8. Non-COI paid editors are widely accepted: say, those paid by public health agencies to improve health info, or academics doing public outreach.
9. "If you don't do anything problematic you could basically go unnoticed for years"; you know, that is actually a perfectly pleasant way to contribute, for many editors.
10. Help:Cheatsheet
11. I'm actually impressed by how little VE messes up the hand-written markup. WYSIWYG HTML editors İ've used were always terrible at that. That said, LaTeX manages better default refs than ":0". Zim-wiki is also well worth trying out; WYSIWYG and WYSIWYM, integrated and working seemlessly on the same Markdown files. I've been meaning to try writing some API-based offline/asynch Wikipedia editor functionality using it for some years now.
12. Levivich, if you haven't read The Mythical Man-Month, I think you'd like it. It details why some things are easier for lone passionate coders.
13. Wikipedia is a democracy, unlike Reddit and most social trap websites. It changes the dynamics. I'm actually a bit uneasy that the WMF's moves towards financial independence from the editing community will hurt that; historically, democracies hang from purse strings.
14. I'd rephrase Levivich's "One Thing=recruitment" to "early retention", once a person has made their first edit, because as I recall that is the stat that's really changed.

But you know, the one thing I could do for retention... I should go do that, I've had it on my mulling-over list for far too long.

Off-topic pre-scriptum: While I don't use "cult" ontologically, I do have a mental category for "religion that mandates (substantial) financial contributions and practices shunning". Fairly objective criteria, and they have a lot of social correlates. HLHJ (talk) 04:04, 31 July 2022 (UTC)

HLHJ, nobody can possibly waste my time on Wikipedia because I am an entirely autonomous volunteer editor and administrator, and comment on what I want to comment on, and ignore what does not interest me. I have been a highly productive editor for 13 years and I do not use any automated tools. I do not need them. I create references using copypasted blank citation templates and I fill in every field manually, copying and pasting the bibliographic information from the sources. I proofread them carefully and use them appropriately. I take great pride in the accuracy of my references and have crafted thousands of references the old fashioned quality control way. I enjoy that work and can create an impeccable reference on my Android smartphone in two or three minutes if I am focused and not talking with my wife, cooking dinner or watching the news on TV. And as I recall, only one of my references has been challenged. There were two American Civil War generals with the same first and last name, and I created a reference about one of them and added to an article that mentioned the other one of them. I immediately corrected my error as soon as it was pointed out. Cullen328 (talk) 04:27, 31 July 2022 (UTC)

Cullen328, I entirely understand that feeling! There is a great deal of satisfaction in careful handcrafting, and that's a lot of why I like markup and WYSIWYM editors. I don't use Huggle or Twinkle; I tried and I just prefer other tasks.

In the elder days of Art,
Builders wrought with greatest care
Each minute and unseen part;
For the Gods see everywhere.

— Henry Wadsworth Longfellow, Wikisource:The Seaside and the Fireside/The Builders

At the same time I tend to feel a certain frustration when I have to do something manually knowing it could be done as well or better by a simple computer program. This is partly because there are way more things that I want to do than time to do them. https://xkcd.com/1319/, but also https://xkcd.com/1205/. I'd never say you have to just paste the URL and let the bot autofill it, but I do say that many editors are less painstaking than you and such features thus improve the quality and quantity of content they write.

It's also a matter of temperment; I often find automatable tasks tedious. Using RefToolbar takes me several seconds to create a reference, and sometimes I chafe at that because I want to get on with writing content. So I guess I feel my time is wasted when I feel obliged to do things a bot script could do; I shouldn't have transferred that sentiment to you, that was unjustified. Some parts of Wikipedia rely on automated tools (for instance, to revert people who replace articles with expletives). In the same way that I prefer computer library catalogues to card-file catalogues, because they are faster and more capable, I find automation makes some things better. And I'm not worried we will ever run out of non-automatable gnomish edits! But on the other hand I've never really programmed on Wikipedia, although I have the skills. I think this is partly because this is recreational, not work, and so, like you, I think less about efficiency and just do things I find pleasant. This mindset can trap our community in bad places, tho, and contributes to the social division between editors and the WMF. HLHJ (talk) 16:43, 31 July 2022 (UTC)

@HLHJ I agree with you that I don't like websites forcing me to do whatever. What I was trying to get at was more "there's no sense of direction" and that might cause unnessecary strife and confusion. I choose to add an article description (a.k.a. Wikipedia:Short description) but I'm given no instructions on what that should look like. I know what I should do, but the average person whose never edited Wikipedia might not. It's difficult to access anything out of mainspace unless you go looking for it... very few new editors are going to search for the Teahouse to ask a question. With the prexisting mobile communication bugs (even though that's got better over time) my understanding is that there's precedents for blocking people because no one can communicate with them and there's no other way to stop whatever distruptive edits they're making. That's a situation that should really not be happening ever. As for social media, I guess the main reason I keep making those kinds of comparisons is that at least for super active outliers like us, Wikipedia inherently has some sort of social element to it. Humans can act... interestingly... in groups. The other thing is that the majority of Internet users use them in some form to read and write, oftentimes on mobile. So making comparisons sometimes could be useful. As for civility, that's a really good point.

In regards to the off-topic stuff, in the interest of fairness, there is no obligatory financial contributions. No tithing. The penny you can spare when you're struggling is worth more to God than the wealthier person whose chalantly throwing around millions like they're nothing. It's something that's repeatedly emphasized as something that makes us different from the "false religions" supposedly. But I definitely think there's a difference between someone who thinks prayer helps them find the meaning of the universe... and what my upbringing was like. An example is that you should never end up in a hospital in a life or death situation needing a blood transfusion and being faced with the "choice" to unrepentantly accept leading to everyone who you've ever cared about shunning you. Clovermoss (talk) 04:40, 31 July 2022 (UTC), edited 10:16, 31 July 2022 (UTC)

I entirely agree that it's hard to navigate the introductory guides to WP. It's a difficult balance, providing enough guidance and simplification to be helpful without providing so much that the user feels trammeled or bored, especially as users vary in temperment. I did see a case in which an editor was blocked for not communicating, but it was some academics running an experiment and they really were breaking rules and norms by dumping contributions written offline by paid students thru a semi-automated account (and the lack of communication was in no way necessary to the experiment, I think they just thought it a waste of time). There may be worse cases. In some ways Wikipedia was a prototypical social network; it pre-dated things like Myspace and foreshadowed their social effects. The social component certainly encourages contributions, because many people find it rewarding. I wish there were better tools for interacting with logged-out readers; the ability to thank at least IPv6 IPs, maybe a link to suggest improvements at the bottom of the mobile view of articles, a simple BlameTool-like way for logged-out readers to thank the authors of a particular passage or edit... I really value comments from non-editing readers, including criticism and statements of incomprehension, they are so helpful.

A lack of obligatory financial contributions certainly removes some potential for abuse of religious faith, though it's not magic; some prosperity churches with optional contributions have had problems ranging all the way into Ponzi schemes. As the Parable of the Ring points out, the ethics of the actions of believers are the measure by which outsiders judge a religious group. Insiders, too; I have yet to meet someone who left a religion without holding the sincere conviction that it was behaving in a way that made it a force against good. This probably also holds for non-religious groups like us. We at Wikipedia get cut a lot of slack because we do good, giving something useful to not only the ~half of people worldwide with internet access, but also to all those who buy print articles in the marketplace or benefit indirectly from the use of Wikipedia by others. We need to be careful not to abuse that. HLHJ (talk) 17:32, 31 July 2022 (UTC)

@HLHJ: The Mythical Man-Month was a very influential book in its day, no doubt, but do you think it still applies with the same force? It was written in 1975, with the most recent edition being from 1995. That's a very long time ago when talking about computer technology. I'm no computer scientist or expert on this topic, but it seems to me that the "communications overhead" was far, far greater 40 or 50 years ago than it is today due to advancement in communications technologies. Programming languages have also changed radically over that time, particularly with the advent of technologies like object-oriented programming, dynamic link libraries, and aspect-oriented programming (AOP). Surely a team of people can more efficiently collaborate in 2022 using Python than was possible decades ago with languages like C or COBOL.

I think the underlying point of the book remains true, but it's less "mythical" than it used to be. I think if we were to graph it, it would be like a curve. For some small projects, the most efficient staffing would be 1 developer, because adding a 2nd dev would just add communication time that exceeded gains in productivity. For example, I recently spent an hour writing a very simple Python script. Had I had help from a second person, it would probably have taken more than an hour, because the time spent communicating would not actually reduce the time spent writing code, because the code is so simple it can just be written quickly by one person. Adding another person-hour would decrease, not increase, efficiency.

At the other end of the spectrum, I still think it's true that if you have 100 devs and you add 1 more, you won't realize any gains in productivity, for the same reason: the time spent bringing the 101st dev up to speed will be greater than any incremental benefit gained from having a 101 devs instead of 100.

But if you have 10 devs and you add 10 more, assuming it's a sufficiently-complex project, I think you do realize gains in productivity. I think it's a "curve" in this sense, with a sort of "slow floor" and "slow ceiling" but nevertheless an incline in between (I'm sure there are math terms for this that I'm ignorant of).

Thanks to things like AOP, modern programming languages are far more extensible than in years past, which means one team of people can work on something today, and a second team of people might use it as part of something else next year. Those two teams don't even need to talk to each other at all, except perhaps via documentation. So in these cases, there can be an addition of devs without any increase in communication overhead at all. That's new, I think, and somewhat lessens the lessons of The Mythical Man-Month. Levivich (talk) 16:18, 31 July 2022 (UTC)

@Levivich: Almost all the technical details are very much obsolete! The basic point that communications overhead can offset the productivity gains of extra workers stands, but I agree there are significant mitigators (partly due to the influence of the book, which was insanely useful for explaining important aspects of programming to the generation of managers who got hit with the need to manage programmers mid-career). I think the mitigating factors are indeed mostly programing and coding practice; people used to quite consciously and deliberately do things that consumed programmer time in order to save now-negligable amounts of machine time. Now, all but the oldest programmers value modularity, abstraction, reusability, quick comprehensability, and all those traits old programmers tended to class as "sources of inefficiency" ("If you wrote it in assembly code you could make it run faster..."). And that thinking is slowly modifying our tools, optimizing them to our values. On the other hand, the basic dynamics of communications are unchanged; writing is slower, and reading is faster, than speech; speech is not asynchronous. HLHJ (talk) 17:47, 31 July 2022 (UTC)

Thank you all for continuing this conversation -- it's interesting to hear all these perspectives. I appreciate the points about "step-by-step interfaces" on mobile -- especially the distinction about car-purchasing being sort of linear, while Wikipedia editing is creative. I need to think about what that might mean for our mobile design. And I definitely appreciate the idea that we don't want our interfaces to coerce users into going further through them than they want to. We should also make sure our designs let users escape, go back, choose a different route, explore -- as they see fit.

One thing that comes to mind is that we did this deliberately while designing "add a link". That workflow guides users through adding a wikilink in a step-by-step process, but community members recommended that we make it possible for the user to discover the fully-featured editing experience if they were curious, or realized that they wanted to do more to the article than just add a link. So we made it possible to toggle to VE or wikitext right from the workflow, and therefore let the user walk through the door to more flexibility if they're ready.

Anyway, this conversation definitely reminds me to keep these principles close to heart.

@HLHJ -- I have a question for you. When you say mobile editing was painfully slow, are you talking about literally the page load times? Or are you talking about the experience being clunky and laborious? MMiller (WMF) (talk) 23:49, 3 August 2022 (UTC)

@MMiller (WMF): Not them, but I can verify that mobile editing is slow, at least for me. I have to wait several seconds when I'm typing or all the text will scramble. Clovermoss (talk) 23:51, 3 August 2022 (UTC)

@Clovermoss -- is that in the Android app? You find yourself typing slowly, one letter at a time? MMiller (WMF) (talk) 23:53, 3 August 2022 (UTC)

@MMiller (WMF): Yes to the app, only sometimes when it comes to the mobile editing in general. I'm a fast typer so it's frustating. Clovermoss (talk) 23:54, 3 August 2022 (UTC)

I tried to provide an example by typing normally on the app just a minute ago but the screen literally froze and caused the app to close. This edit shows what happened [7]. Looks like a test edit but it was literally just me typing actual sentences at a normal speed. Clovermoss (talk) 00:00, 4 August 2022 (UTC)

I'm me , tho I'm sure my experience is no more valuable than Clovermosses's. I had no limits on my typing speed beyond the ergonomics of a soft keyboard, but the process was clunky. It took me quite a few times as long as it would have taken me to do it on a desktop. Fundamentally, the mobile interface was just lower-bandwidth. I couldn't middle-or right-click, cut-and-paste was labourious, I had to cycle thru three soft keyboards to get all the keys I normally have at my fingertips, and using all ten fingers was not on (the Pinephone PDA-style external keyboard would alleviate some of his). Selecting drop-down lists and such was also slow. Spatially, it felt very cramped. Switching windows and tabs isn't so bad, but the screen is just small. I spent a large proportion of the extra time scrolling. Accessing the history or talk tab was slow, and just scrolling to find the edit link took forever. Especially for copyedit-like edits, it would work better with an interface where you swiped right to-edit, center for the preview/unedited article, and right for all the other controls (probably difficult in a browser; I can't use the app). Edit conflicts are painful even with a Desktop; something like Meld (software)'s interface, or just removing those dashes, would make resolving edit conflicts a lot faster. The actual processing is a bit slow, and videos wouldn't play. I found some of the interface popups annoying in that they turned up tardily enough that it wasn't immediately obvious what they were for, and they were novel enough in format that I consciously noticed them. If it's a message I'll see repeatedly, I might come to recognise it, relieving me of reading the text each time; not experienced enough to say if that would happen in time. HLHJ (talk) 00:26, 5 August 2022 (UTC)

@HLHJ and MMiller (WMF): I will say that in regards to browsing and my limited experience with the mobile site, I prefer the app. I don't like everything about the interface but it's nicer for reading in general. Text scramble issues are way more likely to occur when you're editing a page with any substantial content... stuff like the current version of this talk page is guaranteed to mess it up and the subpage I have at mobile editing is at the size where it happens fairly frequently but not 100% of the time. So it seems to be an issue that happens when editing a page somewhere between ~10,000ish and 100,000+ bytes. Usually the text scrambling doesn't look like the previous example but I usually type even slower to avoid it. It's more typically like an actual scramble... different letters from sentences just going all over the place. Again, they look like test edits. Clovermoss (talk) 02:53, 5 August 2022 (UTC)

@MMiller (WMF): It's difficult to express how weird the text scrambling thing is unless you're seeing what's happening as it's happening, but I tried my best to explain it above. I had to retype this edit several times very slowly for it be legible the way it was. If it helps, I could try to find some way to like record my screen when I'm editing text and then email the file to you? I think that'd be okay copyright-wise but maybe I'm wrong. Let me know if it's fine and also if you're interested and maybe also some options for actually doing that in the first place. My mobile edits take place on a Samsung J2 phone, if that helps. My Commons uploads seems to be more specific in regards to the actual model, though [8]. Clovermoss (talk) 04:55, 5 August 2022 (UTC)

@Clovermoss, yes the scrambling issue sounds very strange, and it would be helpful to have a video to pass along to the team that works on that part of the experience. I believe some Android phones have native screen recording or you could find an app. We at WMF frequently exchange screen captures to show bugs or designs. You can email it to me, or even upload it to Commons (if you don't capture any personal information in the video). My email address is mmiller@wikimedia.org. Thank you! MMiller (WMF) (talk) 18:20, 5 August 2022 (UTC)

@MMiller (WMF) Okay. I'll give it a go and keep track of the revelant timestamps. Clovermoss (talk) 21:22, 5 August 2022 (UTC)

@MMiller (WMF) I've sent a ~7 minute video to you through email that shows one method of me trying to edit this talk page that causes the text scramble issue. I tried to write like 2 sentences over the course of that time and a lot went wrong. As I said, hard to explain unless you're actually experiencing it. Hopefully it's helpful. I'm still open to my offer to do other things as a frame of reference if you think it'd be helpful to see where I've identified issues. Clovermoss (talk) 01:34, 6 August 2022 (UTC)

@Clovermoss -- thank you for sending the video! I can see how that very slow and scrambly experience is frustrating. I forwarded it to @JTanner (WMF), who is the product manager for the Android app. I also showed her your page of mobile editing thoughts. She may be able to get some perspective from her team on what you're seeing. MMiller (WMF) (talk) 19:44, 8 August 2022 (UTC)

Thanks! I appreciate you sending it to the right person. I admit I'm not super familiar with whose in charge of what. Is there some way to figure that out other than trial and error? Like a page that lists all WMF staff or something? I hope my thoughts are useful. I will admit I'm still messing around with the mobile app editing interface but I think some of my initial impressions are important.

@JTanner (WMF): I would say that something that's vital, apart from the text scrambling issue, is the inability to copy links to diffs. Diffs are important because if you're trying to discuss an edit on a talk page or even just in general, people will want to see a diff to know what you're referring to. Clovermoss (talk) 19:49, 8 August 2022 (UTC)

Hey there @Clovermoss, thanks for taking the time to use the app and write such detailed feedback. It is very helpful and valued.

At a meta level, to find who works on what, a good place to visit is the Product Department MediaWiki page where it says product teams. You can comment on talk pages for that team and someone should get back to you. The Android team's project page can give you a good idea of what projects we worked on in the past, and what we are actively working on. We make it a priority to also work on tickets that come into our Phabricator Board filed by helpful folks like you when we have capacity. However, our team always makes it a point to give an idea of the "when we have capacity" will be. I share this so if you have feature ideas or notice bugs, outside of what you've listed here and want us to triage it within a week, that is the fastest way to grab our attention. However, we do our best as well to keep an eye out on talk pages or when we get pinged on helpful pages like yours. I have some time scheduled with our team's engineer who has been on the team longer (I have been on this team for 2 years), and can help me better understand why things that predate me may not have been prioritized previously so that I return to this talk page beginning of next week and provide some context for why things are the way they are, and more importantly what plans we have to address it in the future (and an estimate of how far in the future).

Zooming in specifically about the scrambling issue, it is a very unpleasant experience that has to do with syntax highlighting for devices like yours. What is happening there is the system freezing momentarily and causing keystrokes to queue up in the background, so when it unfreezes the keystrokes fall out at once, and is particularly prevalent when trying to edit a full page due to the amount of content that is there. Although with section editing depending on the type of device and amount of content in that section, it can be somewhat sluggish as well. For a while we disabled the ability to edit a full page and only allowed section editing, to limit performance issues. Recently we enabled full page editing because people (understandably) wanted to edit text that sometimes exists above a section, in a template for example, and the only way to allow that was to enable full page editing. The app was originally created for reading, in the past few years editing capabilities have been added based on requests we've received and our desire to serve mobile app users more holistically. That has come with a number of challenges, that I am very excited for us to work on in collaboration with people like you. Specifically concerning fixing the syntax highlighting issue, I can tell based on the engineer in 2017 abandoning the ticket it is quite a large task, but I will have a better understanding by the end of this week of just how heavy a lift it will be to give a better idea of when to revisit it, that is, if fixing this issue specifically is even appropriate. Our team is investigating somewhat showing the mobile web editor by building a layer over it, which would fix the syntax highlighting issue as well as other issues, while also keeping some of the unique features of our editor like Dark Mode. It is a very big project but something being prioritized because ensuring people can have a holistic experience in the app is very important to us. In the interim, any changes we can make to improve the existing native editor like link preview or easier ways to add an image, we are making based on requests that have come in, so it is still helpful to point these things out.

When thinking about copying a link to a diff are you thinking about what happens when you :

• Visit an article
• Click the top three dots (Overflow menu)
• Click Edit History
• Click a diff
• Click Overflow menu
• Click Share Edit

If this isn't what you had in mind, or it feels unexpected, I am happy to hear any improvements or additions you may have in mind!

I will keep an eye on this page and watch it. I also will provide an update beginning of next week regarding some of the other things you shared. Thank you again for taking the time to do this! JTanner (WMF) (talk) 22:50, 8 August 2022 (UTC)

Fortunatly, I had nothing like that sort of text-scrambling! I've seen newcomers write edits that are scrambled like that, too, maybe that's why. I'd assumed test edits or pocket edits. An increasing diversity of editing interfaces does create a lot of WP:THEYCANTHEARYOU-style social problems where you just don't know where other editors are coming from; theory of mind and teaching gets much more difficult.

For me on the website, everything seems to work as-designed (even the videos fail by displaying widely-temporally-separated still images, and the odd flicker of motion). I guess I was editing the website on mobile Firefox and thus taking advantage of a lot of older designing and testing. Better phone I/O would fix for me, but while waiting on that, I'd like more-accessible interface-navigation links on the mobile site.

Aren't apps generally higher-maintenance than websites? HLHJ (talk) 23:23, 6 August 2022 (UTC)

Evaluating semi-automated first interactions

@MMiller (WMF), are there any stats on retention rates as a function of which templates are posted to a new user's talk page? If some uw templates have higher retention rates than others for extrinsic reasons (like they fail to explain to the editor what they need to do or come off as plain rude), then fixing them would be a very low-effort way of improving retention. If the data exists, I'll volunteer to code the templates by traits and run some analysis. Broader stats on what semi-automated first interactions are most helpful and repellent would also be useful. HLHJ (talk) 20:34, 7 August 2022 (UTC)

@HLHJ: I'm not MMiller, but I'd suggest maybe asking Sdkb some of these questions? They've done a lot in regards to streamlining welcome templates to make them more useful and less overwhelming to new editors, so if anyone's aware of stats regarding this, they'd be the first person I'd think of who might know. Clovermoss (talk) 16:40, 8 August 2022 (UTC)

Mobile app editing example

Options to share a diff on the android Wikipedia app, there's no way to share a diff normally. Many other mobile apps, such as YouTube will have some sort of "copy link" feature. For the Wikipedia app to be missing a feature like this is, in my opinion, a glaring omission that should be remedied.

@JTanner (WMF): Thank you for getting back to me. Your confusion about what exactly I was referring to, is in my mind, a good example about what I mean about how can there can be a disconnect between the WMF and the broader community of editors. So the issue isn't that I can't see diffs, it's that I can't copy the links to them. Ask any experienced editor on why this is important, like Levivich, and you'll get lengthy answers, I'm sure. The gist is that it's important for communication. If you want to start a talk page discussion because you were reverted on an edit you made to an article, it's standard to link a diff to the changes you tried to make. If you ever start a thread at WP:ANI, you're expected to provide diffs. They're very important. This [9] is an example of what I want to be able to do. As for the problem being my phone, I was able to type all this on the desktop version of the app just fine without that issue, so I'm disinclined to believe that's the reason. Or maybe the desktop version of the site has already figured out how to mitigate that response? But I did want to say that I appreciate your reply. I think more open communication can do a lot of good and I know you're trying to make the mobile experience better. Clovermoss (talk) 23:28, 8 August 2022 (UTC), edited 23:45, 8 August 2022 (UTC)

Thanks @Clovermoss: for taking the time to share an image. Being able to share diffs is something we deliberately put in because we do really understand the importance of being able to do so! Believe it or not, quite a few members of our team are long time editors. Below is a screenshot of a Pixel 5 where it does provide the copy link option.

Options to share a diff on the android Wikipedia app, copy link option on a Pixel 5 on version 2.7.50417-r-2022-08-02.

Now while some of us are editors we do have different devices. We have recently requested to get some older and lower tech devices so it can expose variances like this that often happens with Android devices because it can vary. So thanks for sharing that on your device the copy option is not showing up we can investigate why that is happening.

The Desktop experience uses different back end technology than Mobile Web and the apps. When Wikipedia was created it wasn't created with mobile in mind. As you've correctly pointed out, we are playing catch up with some things on mobile and how that manifests is performance issues with the editor, which we are working to fix. However, the reason you don't have editing issues when editing on your mobile device using Desktop is because it is using different APIs and the apps were originally built for reading. That is clear in the interface and how much better the apps are when it comes to accessibility for low vision users for reading. Hopefully in time as we built a great reading experience the editing experience will catch up. I also completely agree when it comes to open communication and hope we can continue this dialogue as we sort through enhacing the mobile experience. JTanner (WMF) (talk) 00:20, 9 August 2022 (UTC)

@JTanner (WMF): I'm glad that this is something that your team has thought about and that you'll look into the issue. So your confusion was because it should show up as an option for me. Honestly one of the best case scenarios there. I'm aware that a lot of WMF staff do edit Wikipedia, I actually pointed that out earlier in this thread. My understanding is that WMF staff are encouraged to edit, yes? Which makes sense because how can you understand what people are going through if you have no experience doing it yourself? I will say a lot of active editors have became disillusioned with WMF involvement over time and that I guess some of that has rubbed off on me, even if I'm generally more hopeful that things can get better. I still do think that the app could use some susbtantial improvements in certain areas and I look forward to your more detailed response in the future since you said you're looking into that. Clovermoss (talk) 00:28, 9 August 2022 (UTC)

@JTanner (WMF): Just following up since it's been a week. Do you need more time? I'd just prefer to know that it's still something you're thinking about, if it's still on your mind. Clovermoss (talk) 19:41, 15 August 2022 (UTC)

Hi @Clovermoss! Yes this is still top of mind for me, I am working through this and planned to provide an update tomorrow. See an example here of me filing some tasks for some of the feedback. I planned to share the corresponding links and a holistic update of what to expect and how we can continue to engage on the various topics, I appreciate your patience in the meantime.

In respect to feeling disillusioned with WMF, I can understand that, I am very big on being honest about what can and can't be delivered and context around decision making. I also strongly believe in open collaboration and, I hope that transparency can help you at least see the commitment from Android. I believe the Growth team and some others have a similar approach, though I can't speak for everyone. Also, you are 100% correct, that the confusion was that for the 10 devices we tested on the link and copy option in the bottom sheet did appear so I thought it would show up for you, seeing that it didn't is helpful to know that the link showing up is device dependent and unexpected behavior. I hope that makes sense!

But to confirm, this is still top of mind, and progress is being made on sharing a response.Thanks again for your patience! JTanner (WMF) (talk) 00:17, 16 August 2022 (UTC)

@JTanner (WMF) Thanks for getting back to me. I appreciate that you're trying to improve some of what I've brought up, I look forward to hearing the further update you're working on. I think the recent comment I wrote about remembering that real people are behind the screen working on these things and want them to go right too is something that's important to keep in mind.

Here's a paragraph break to give a visual example of something I believe I brought up? I'm not sure if it's on the version of the mobile editing page you looked at (I keep updating it over time, it's definitely something that's a work in progress). The reply tool on the browser doesn't do that. I know you mentioned that the API for these things is different, does that mean that comparing how the two work isn't really that useful? In regards to the testing on ten devices thing... I seriously have no experience with software development, so maybe take my opinion with a grain of salt here... but that seems low to me? Again, I'm not sure what's typical and I'm sure trying to figure out something that works with the numerous mobile devices that exist today is nuanced and time intentsive (so many things to account for!) but to a laywoman 10 just seems like a limited number. Maybe expanding it to a greater variety would help with identifying potential issues?

In regards to greater community involvement, I'm all for that. I seriously think a lot of issues kind of get worse because people feel like they're not being heard and that leads to them feeling like their opinions don't matter. Then there's also the action or inaction that takes place after that. If it's all words but nothing comes of it, people end up with more resentment over time. A lot of experienced editors have invested hundreds of hours into Wikipedia, so I think it's understandable that people would feel strongly about their involvement here. Life is complicated. Obviously you're not in control of everything the WMF does, but I appreciate that you're taking the time to listen and to some extent, validate my experiences. That's something I think is important in day to day life, y'know? Clovermoss (talk) 01:23, 16 August 2022 (UTC)

@Clovermoss

Thanks again for putting so much effort into thoroughly testing out the app, and being patient with me as I went through the feedback. I needed to check in with some of the team on why certain features are the way they are because some of the decision making was made prior to me joining the team, and I’ve found that some flaws or quirks about the app exist because of the limitations of our technology. I wanted to be realistic about what could be done and honest about what would be a little bit out of my control but share ways it can be advocated for. I appreciate your note that we are all real people trying to do the right thing, it is very true! It is also true that it isn’t great for folks we collaborate with to feel unheard. Throughout our development we absolutely must partner with as many people with diverse use cases as we can to build accordingly. Our community relations specialist @ARamadan-WMF who I want to loop in, has really been helping us take our previous process and improve it even more, by adding the aspect of responding within a day or a few days (depending on the complexity of the challenge) whereas we’d previously respond within a few days to a week, based on team capacity. She also will help surface feedback that is common between the iOS and Android apps. This is important because it helps us see what new or recurring issues are impacting people on both apps, which would point to an API issue. When we get common requests about things related to APIs we can share the sense of urgency with the teams that work on APIs. Hopefully me sharing this also answers if it is helpful to compare platforms. It is helpful to hear expected behavior, especially as a cross-platform user, even if we know the issue exists and our team specifically can’t address it right away, when the teams that work on APIs are prioritizing support (which extends beyond the apps teams) we can make the case for getting our bugs and features prioritized. Unfortunately because we aren’t like tech giants each team doesn’t have dedicated engineers that work on APIs and because Wikipedia was originally built for Desktop we are reverse engineering some things here the best way we can with the added complexity that not all workflows should be just shrunken down to a mobile device. So while our team specifically can’t fix everything (and some workflows should be mobile first), it is helpful to see what common concerns arise, and escalate and prioritize it accordingly.

You can get an idea of what our team is working on currently via our Phabricator board. You’ll see some of the things you’ve raised are there, along with some of our bigger projects that have been long requested, and take a bit more work to get done. I also encourage you to file tickets for any feature requests you have or bugs you notice and tag our board. Amal will keep an eye on this page, but we triage tasks created and tagged with our board twice a week so it is the fastest way to grab our attention. It is also the best way for you to track the tickets by subscribing, collaborating with us on implementation and see possible progress! For example, volunteer SD0001 created this task requesting improvements to cross platform communication tools, and you can see platform by platform how things were, or are still being addressed varied and the collaboration process throughout. Sometimes volunteer developers and designers will even work with us on Phabricator. It is also a good way to see our test coverage. To your question about 10 devices, this is actually above average in some cases, we also include automated testing and release it on a Beta of the app before it goes to the production version. The Beta version has over one million downloads who report issues.  A good example ticket to see how all of what I mentioned above works together is task T312667. Issues with horizontal scroll were repeatedly raised on both iOS and Android, so our teams filed a ticket to request it be fixed by one of our backend teams. TheDJ is a volunteer developer that gave helpful input along the way. AborbaWMF is our test engineer that tests the bug on both platforms, along with using automation and partnering with neurodiverse organizations that specialize in things like unit testing and smoke tests before we release it in Beta for further testing. We often do additional usability testing in multiple languages with different experience levels and technical fluency to surface anything we’ve may have missed, but with Android in particular being so device variant, as you pointed out there can still be things we miss, so it is helpful to get feedback from people with diverse devices, geographies, wiki experience, languages, technical fluency and use cases, like yours!

Now to zoom in a bit on some of the feature ideas and bug findings:

• In 2022, it automatically downloaded articles I had saved for offline reading in 2019. This was about ~60 mb worth of space. I was taken aback by this a bit because the last time I used the android app it was on an entirely different cellphone years ago.

Yes, this is the syncing of our Reading Lists feature. When the feature was initially released, our research determined it was preferred to have offline reading (which means downloading the article) on by default. If you go to Settings you can turn off download reading lists articles and turn off syncing across devices. That should save you some storage. 60 MB storage size is fairly below average, we do keep an eye on our app size to make sure it doesn’t become too bloated.

• I tried to see if I could create this page through the app, it let me search existing subpages of my userpage but as far as I can tell would not let me see/create a page that did not yet exist. Once I had created this page in desktop view on Chrome, the page automatically loaded because I had previously searched for it.

We do have a task for this that has been around for quite some time (T131284). We could quite literally change the code today and allow being able to create a page, it is a feature I desperately want in the app myself, but the experience would be horrible at the moment. Because people can’t do things like add templates (amongst other things), creating a page and reading a page that was created in the apps would be a really bad experience and likely requested to be turned off so before we can enable this we really need to fix some infrastructural issues, which includes requesting changes on the back end. This is due to the app being originally created (quite a while ago) for reading and over time after requests for editing capabilities we’ve been doing our best to reverse engineer it to allow for some short form editing, but things like creating a page from scratch, we must continue the work we are doing right now (T307923)  on investigating possible infrastructural changes that would unlock more editing capabilities before we can tackle something like this. The app needs to build everything, and can’t just copy the code from the web version, and with only a small number of developers we don’t have all the features we want to have simply because it takes so many years to build them. Wikipedia has had more than twenty years and a big number of developers focused on making a superior desktop experience to get to where it is on the web.

• It keeps track of literally every link I ever click unless I use the back arrow a million times. I don't get what the point of this is when there's the option of opening new tabs? Maybe to mimic a web browser? To be completely honest, not a fan. Feels unnecessarily clunky and might give the app issues longer-term even if I was just using it as a reader. Maybe clearing my cache regularly would help? I have found a way to clear my browsing history and tabs, which is slightly better. Update: changing my browsing habits when reading Wikipedia has helped a bit. I open a new tab each time I click a link so I don't have to click the back arrow a million times to get back to the main page/look through my recent contributions.

The back button has been something we’ve received a lot of feedback about via the app store. Some people share your view, some people have the complete opposite opinion. What we’ve done to address the different preferences at the moment is release of a feature called Customizable Toolbar. If you want to return to the Explore feed (Home) quickly you can update your toolbar to put the explore action there. Otherwise Explore is always accessible in the overflow menu. If you didn’t know about the feature that is another point. A major project we are working on right now is called “Onboarding and Guidance” . We are doing a deeper dive into user preferences and motivations to evaluate the architecture of the app, and where we guide people to learn about features in the app when based on preferences. There may be modifications we make based on what we hear from folks using the Android app. So far we’ve conducted an in-app survey and will continue community engagement on MediaWiki, so I hope you will continue to engage with us as we share what we’ve learned and some ideas to improve feature discovery.

• Italics function interestingly. Not sure if it's just a VisualEditor thing or the app itself, but when I clicked the italics button at the bottom it added the ' that I would use in source but it also italicized the words as they appear on my screen?

This is actually the use of syntax highlighting and perhaps we can better explain this in the app, Task filed (T315332) .

• Although as far as I can tell, it's actually impossible to use VE on the app, which seems like a weird omission. If there is a way to switch it from source, it isn't glaringly obvious. Maybe I'll figure it out later.

This is the same point as above about investigating possible improvements to the editor. It is a heavy lift, so I do want to be transparent that we are still very much in the investigation phase but certainly noted, and something we have top of mind. For now we are working to improve the native editor in ways within our control, so please keep an eye on (T313223 ) and if there are other ideas of enhancements feel free to drop it in the task, and test it in Beta once we are a little further along in a week or two.

• Default edit summaries are "fixed typo", "fixed grammar", "fixed links" or "other". I'm someone whose always left relatively detailed edit summaries so it's slightly annoying to have to press "other" each time when I can just type the edit summary right off the bat the way I'm used to editing. I do like that you automatically get to preview how your changes look before the edit is published, though.

Completely agree! We are working on this improvement (T296952). Something we are weighing is dropping the other option and allowing folks to just drop in the summary while still showing “fixed typo”, “fixed grammar”, etc. as options. The tradeoff is the “other” option allows users to save common phrases they use, and getting rid of “other” would remove their custom saved answers. There may be a middle ground so we are talking through an implementation that doesn’t feel awkward.

• It tried to prompt me to add an "article description" to my user subpage. As far as I can tell, seems to be prompting me to add a short description, but that should not be happening outside of mainspace and the app itself should be able to tell that a userpage is not an article?

Fair point. I’ve filed a task to disable the call to action for non-articles (T315237)

• Anytime I click anything, I have to click "read article". Whether that's an actual article, talk pages, user pages, etc. Having the box say "read page" would be more accurate.

It appears this is talking about link preview, I’ve filed a task (T315239) to change it to just “read”, we’ve gone back and forth between article and pages (specifically in English). Although, I will need to park the task so that our team can talk about this a bit more because there are other places where it says read page or read article, where read alone wouldn’t make sense, and page wouldn’t work uniformly, so this task will require having a better idea of the different cases where “Read Article” is used and if we should build in logic.

• When reading Clover Bar (provincial electoral district) it offered to give me "quick facts"? Clicking that seemed to give me an infobox, but maybe it's some other feature I'm not aware of. Nope, it's definitely an infobox now that I'm looking at it on my computer. This would be the second time I've noticed something being called what it isn't referred to by the wider community. I think that this could cause unnessecary confusion. If you ask for help adding "quick facts" to an article, the assumption for the average editor would be something along the lines of adding content/citing a source not how do I add something to this infobox?

This type of change would need to take place on iOS and Android. Infobox was one of the components used to create the “quick facts” in the app, but there may be a way to create parity. I’ll have more discussion with the teams to see the implications to change it but it is fair that it could cause confusion amongst cross platform users.  (T315335)

• User pages are called "user profile pages" when you use a drop-down menu to click on them.

Agreed, we’ve added it to our board (T315242)

• As far as I can tell, there's no way to edit a page in its entirety... you have to go section by section. Despite this, it's impossible to edit subsections directly, which makes it very difficult to edit longer articles. Update: It is possible to edit an entire page at once, although it wasn't immediately obvious to me. This does not work that well in long discussions as your text just kind of gets scrambled all over the screen unless you wait like a second or so before pressing every letter. It's essentially guarenteed to happen on large pages, but it can also happen on relatively small ones. It still looks like it's impossible to edit subsections directly, though. It's either the entire page in source or sections by themselves in the standard editing method available through the app.

As I mentioned earlier, this has to do with Syntax Highlighting (T164936), which our lead engineer investigated as his hackathon project during Wikimania. He was already looking at this code based on the editing improvements we were making (T313223) and he made a lot of progress that should greatly improve this problem. I recommend subscribing to T164936 so you can see our progress and give us feedback once it is available on Beta within a few weeks.

• It's impossible to mark an edit as minor.

Yes, we have this task on the board (T296952)

• Redlinks show up as blue links, even in preview. So you don't know if you're making a redlink until you've already hit publish.

Here is one of those examples of needing support from a backend team for the fix. We have filed the ticket (/T315347)

• There's a box that claims that my edit quality is perfect, whatever that means. There's no provided explanation. I think this has the potential to be flawed... how can an app know if I'm doing things right? Also, no one's perfect. If someone's edits are deemed problematic by others, though, this whole "but I have a perfect edit quality" thing might not be the best idea. There's a few other things listed on this "panel". I have a box that says "0 views" which I'm assuming has something to do with my contributions. There's also an "edit streak" with a # of days. I do not like this feature because sometimes I take wikobreaks and I also don't want to feel obligated to edit Wikipedia every day just to maintain a streak even if I edit fairly frequently. It seems to only count edits made on the app itself as part of my ongoing streak since it's reset a bunch of times.

This is a feature that was rolled out a while ago that our team has interest in revisiting. The Growth team is actually working on an impact module that is quite similar and is doing a lot of research on it. We plan to work closely with them to learn what works and what doesn’t and will make adjustments to these features based on the Growth team’s outcomes. Hopefully waiting to see their implementation will also create parity.

• I got a notification in the app for a Commons reply that took me to the desktop version of that site in my browser? I'm very confused about why all these keep switching... shouldn't everything be consistent to whatever I'm using in the moment?

This is true, but unfortunately not much we can do about it. Our app is a handler for “wikipedia.org” links, but Commons is under the “wikimedia.org” domain. But we still receive notifications natively from Commons, because it’s under the Unified login umbrella. The same applies to Wikidata.

• Text changes in diffs show up as green. The colours used in desktop view are yellow and blue.

This is true, I’ve created a task for our team discuss it with the iOS team /T315348

• Looking through my contributions does not tell me if my edit is the current version of the page.

I’ve filed a task for our team to discuss this a bit more (T315380). If you have ideas of what was expected, feel free to drop it in the task.

• I have discovered a few days after using this that there is a way to access my watchlist. It is not very obvious at first. You have to click "more" near a box listing your global contributions and then it's an option there.

We’ve gotten complaints from readers when editing features are too prominent. We hope with the onboarding and guidance project it will help us understand if watchlist should be more prominent when someone indicates they are using the app primarily to edit or when we should show users the location of the feature in their editing journey.

• I have the option to "subscribe" to pages in certain contexts that I did not notice before? I can subscribe to my own talk page, for example. Still trying to figure out what this does, but I'm assuming it's meant to give me automatic notifications on changes for pages that otherwise wouldn't... a watchlist that's seperate from my actual watchlist? Update: my assumption was correct, except that subscribing to a thread gives me a phone notification when anything changes. This works best with talk pages where I only care about replies in one section and I'm not sure if someone will ping me.

Topic subscription is a new cross platform feature that was led by the Editing team. They’ve created some helpful documentation about it. If you felt there should be more onboarding related to this feature, or there is unexpected behavior, feel free to request feature changes you have in mind and tag our board as well as Editing!

• There seems to be an issue with longish text overlapping (e.g. lists). Creating an extra line break and then getting rid of it seems to fix the issue.

Would you mind sharing a screenshot of this and tagging us? We tried to recreate what you were describing but weren’t able to, sorry about that.

• When replying to someone on a talk page, you automatically ping them. I like this feature a lot but it's redudant when it pings someone on their own talk page.

Good idea:T315381

• When looking at diffs on my watchlist, my options under the text changes are "thanks" and "rollback". If I click the three dots at the top, I can find the option to undo an edit instead. I do not like this setup. I wonder how it would appear if someone did not have the rollback userright. The option to "thank" also shows up in diffs where IP editors have made the edit. I tried to thank an IP even though I know it's impossible just to see if it would work. It does not, I got an error message saying that I did not have a valid recipient.

Good point for IP Editor thanking:.T315382

When someone doesn’t have rollback privileges they see the undo button in place of rollback. The overflow menu would only contain watch and share. We put undo under the overflow menu for rollbackers to not have a clash where someone wanted to click rollback and they hit undo instead or vice versa. We also were taking into consideration the limited real estate for smaller screens. If you have other ideas feel free to make a feature request and tag us!

• When participating in an AfD, you can't reply to people like you can on talk pages. You have to edit the page in source. For some reason, using the four tildes does not show my signature in the preview before my edit is published. It was super weird to just see four tildes but otherwise normal preview functionality.

This is because AfD discussions are not in the Talk: namespace (they’re in the Wikipedia: namespace), and therefore we don’t handle them in our native Talk interface. We could make a special case for AfD pages, but it would get very hacky for multilingual support, and I’m not sure whether the DiscussionTools API supports pages outside the Talk: namespace, either. To get an idea of this I created a spike to understand the severity of trying to do something like this (T315385).

• As far as I can tell, there's no easy way to be able to get links of diffs to show other people. Clicking "share" at the top of a diff gives me the option to email or text the edit to people.

After further investigation this seems to be a device-specific issue. Nevertheless, our intention is for everyone to be able to share diffs so I’ve made a task (/T315387)

• App pageviews don't count as "seen" when it comes to my watchlist. I was able to see "new" changes in my watchlist on desktop for my talk page (including my own replies) that I had already seen. It was a weird experience.

Looking at your talk page natively in the app does not count as a pageview, since we’re using the DiscussionTools API to get the contents of the conversations. We will talk to the Editing team to see if  querying the DiscussionTools API should count as a pageview.

• There's no obvious links to instructions anywhere. No links to the Teahouse, nothing explaining the basics of how you would even edit, etc. I think this lack of any sense of direction is a disservice to new editors because it prevents from being able to look before they leap.

There is some onboarding for app specific editing, for example if you click the edits tab we have documentation about suggested edits and other helper links for each one. Unfortunately we can’t simply reuse all editing documentation that exists because a lot of the instructions are pretty desktop-centric. In regards to the main editor, we could better guide people and inform them of policies, the Editing team is working on a feature that integrates policy into editing. We hope to learn from them, especially around multilingual implementation and also see what things we can improve in our onboarding and guidance project. There is also the point about connecting users with other humans for help, which is why we spent the last 12 months enhancing communication related features. There is room for improvement, but we will first evaluate the outcomes of talk page improvements, and appending namespaces like help to our search.

JT: We have the help namespace, and we are working on onboarding. I think we have to create instructions that makes sense for our context

• Barnstars don't display through the app on talk pages. What you see is a bolded version of the title of the specific barnstar you're awarded (e.g. The Original Barnstar). Images display fine in articles themselves, but not on talk pages.

Native presentation of talk pages allows for a very limited subset of html formatting, and will not support something as fancy as barnstars (or custom-formatted signatures, etc) at this time. However, as we focus on infrastructure improvements, we will certainly revisit this! JTanner (WMF) (talk) 22:55, 16 August 2022 (UTC)

@JTanner (WMF): Thank you for the very in-depth feedback, it was worth waiting for :) As for the longish lists, I took a screenshot of what it looked like one of the times it happened, when I upload it I'll ping you. I'll try to figure out Phabricator more since it seems vital to how your team operates. As for help pages, I don't think all of them are desktop-specific. Even if you can access the help namespace if you type in exactly what you're looking for... I don't think the average new editor is going to know how to do that. I think something along the lines of Help:Introduction to Wikipedia would be helpful (which you can see on the main page in desktop under "anyone can edit"; isn't really glaringly obvious, again because banner blindess, but it's still better than what I've been able to find). At best, this would look like something like a Help? option in the sidebar or something like the help button that's at the top of this page but in a more mobile-friendly format. I think it's really important that there's some obvious/initutive way to access resources like this. As far as I can tell, the Teahouse actually seems to function quite well on mobile, despite it being a huge page. I think it would be really important to show this somewhere because understanding is that there's a demonstrated editor retention benefit in telling people about the Teahouse. I might be a bit biased as an adoptee of Nick Moyes, but still, there's a lot of people you could listen to about how great the Teahouse is. Clovermoss (talk) 00:15, 17 August 2022 (UTC)

@Clovermoss: happy to read it was worth the wait, and excited about the prospect of you engaging with us on Phabricator!

That is fair, that we should have better guidance for newcomers around editing. I will make it a requirement to figure out for our onboarding project. We will explore existing options like Teahouse and the Introduction to Wikipedia page and specifically test exposure to these resources to see how intuitive or helpful it is for newbies in the app context or if an alternative should exist. JTanner (WMF) (talk) 00:37, 17 August 2022 (UTC)

@JTanner (WMF): Okay, I'm glad to hear that you're thinking about the newbie experience. I also just wanted to say that, in general, if there's anything you ever want me to try out or my opinion on something, feel free to reach out whenever. Clovermoss (talk) 01:01, 17 August 2022 (UTC), edited 11:42, 17 August 2022 (UTC)~

@Clovermoss This is a thread and a half! You pinged me, so a quick reply is probably appropriate, as I really don't have time to read through it all. For JTanner (WMF)'s benefit, the Teahouse on en-wiki is extremely supportive and welcoming to new editors and their needs (our prime purpose!). I suspect few of us are that familiar with WP:VE, and none of my fellow hosts (as far as I am aware) use or recommend the mobile app. I am aware that many practical guides assume Sourvce Editor is being used, whereas the reality is perhaps different.

I installed the mobile app on my trusty iPhone5S a year or so ago, but soon removed it as it offered me nothing over direct browser access in Safari. I read and edit a lot on mobile - but nearly always in Desktop mode. I do my admin work on a desktop as I don't find that level of detailed work is best suited to mobile. Cullen328 would disagree, I'm sure! I'm a bit of a dinosaur and have never managed to understand how Phabricator works (and can't even find a way to update my username there). As a trainer, I know we're now pushing new users towards VE, so it's incumbent on me to use that more - which I've been doing since earlier this year. Unfortunately, I see no reason to install an app for reading or editing Wikipedia, so I'm unlikely to want to spend time assessing it without good cause. Not sure if these comments are at all relevant, but there you go! Regards, Nick Moyes (talk) 01:12, 17 August 2022 (UTC)

@Nick Moyes: Yeah, this discussion has definitely been going on forever, especially when you consider that this is actually 2 discussions after the original one! There's been a lot of different tangents... some stuff about the new editor experience, some stuff about my general life experiences, some stuff about mobile editing, etc. I'll likely make a more concise reflection/summarization at some point. I've definitely become a wall of text person, that's for sure. Feel free to take your time reading through all this if you're interested. As for the app, I'm not trying to convert more people to using it, but I was interested in learning more about what editing might look like from different perspectives because generalizing my experience as universal in general isn't the best idea, y'know? I still prefer source myself, but that's another reason I'm interested in learning more about VE. It is slightly exciting tackling something that makes my head spin a bit and actually being actively involved in some changes, though. I feel like I'm actually contributing to something that matters, even if that's a reason I've kind of always been interested in editing here. But there's a different feel to it at the moment, I guess. Clovermoss (talk) 01:26, 17 August 2022 (UTC)

If I get a moment, I'll try! Learning new skills so as to help others is important. But for me at the Teahouse, I see no reason to recommend downloading the mobile app from Google Play store (nor to familiarise myself with it after my first foray with the iOS version). I would simply advise newcomers to edit on the mobile browser, either in mobile or desktop view, and actively steer them away from downloading an unnecessary app that none of the helpers use and no one needs. I'm afraid I see nothing at WP:MOBILE that makes me feel WMF should be spending time and resources on this when the browser generally works fine in mobile or desktop mode, and we should focus our attentions on improving those. Nick Moyes (talk) 12:39, 17 August 2022 (UTC)

@Nick Moyes I think the app is mostly intended to be used more for people who are interested in reading Wikipedia, not editing it. I wouldn't suggest people use it if they don't want to... there's a lot you can't do and my opinion is more "make it better since it exists and for the people who do want to use it, maybe if the editing experience is more initutive they'll give it a try on desktop or regular mobile view". First impressions can leave lasting impressions. Since it exists and there's dedicated WMF staff whose jobs are to improve it, I figured I might as well try to help make it better. Especially since no one in the wider community really focuses on it for the reasons you stated. There's nothing that really explains how you even edit anything other than the suggested edits and the actual edit buttons, so I think that offering some sort of guidance to people to try and figure things out other than nothing and hoping for the best for the people who do would be a good idea. I've always valued your opinion, by the way :) If you don't want to recieve further pings here let me know and I'll respect that. Clovermoss (talk) 18:14, 17 August 2022 (UTC)

No, I'm always quite happy to offer comments, though can't get too deeply involved in complex, long-term stuff right at the moment. I do quite a few UX testing sessions as a side hussle, but they're all short-lived feedback sessions, not long development projects like the app must be. I appreciate it's intended as a reading tool; I just didn't see what was different about it. (My one gripe with mobile view in a browser is that you can expand and collapse a section at the top of each sectopn, but there's no 'collapse' button at the bottom of each section to close it again without having to scroll all the way back up. So pages with long sections just keep getting longer and longer and harder to navigate). Nick Moyes (talk) 18:37, 17 August 2022 (UTC)

@JTanner (WMF): I also wanted to clarify that it wasn't the app that took 60 mb of space but the automatically downloaded offline articles. According to my phone settings, the overall size of the app takes 427 mb. I do like that I was able to transfer it to the SD card on my phone, otherwise I wouldn't have the space for it. But you pointed out that my storage space was below average and I was wondering if it wad the 60 mb figure you were considering in regards to that. Clovermoss (talk) 05:44, 20 August 2022 (UTC)

@JTanner (WMF): By the way, here's the file demonstrating the overlapping text issue: File:Overlapping text example on the Android Wikipedia app.jpg. Any ideas what's happening there? Clovermoss (talk) 12:10, 22 August 2022 (UTC)

Hey there @Clovermoss I shared the overlapping text issues with our engineers, and we believe it will be fixed within the next two weeks (pending passing testing of course) with task T164936. I've subscribed you to it so you should get updates as it advances. Once it is on Beta, I'll ping you so you can have a look and let us know if it fixes the issues on your end. JTanner (WMF) (talk) 16:42, 22 August 2022 (UTC)

@JTanner (WMF): Thanks! Also thanks for linking the guide about Phabricator on your talk page, I'll check it out. I've always kind of known of its existance, but I've never really tried to do anything like file tasks myself. I really am interested in looking into it. Clovermoss (talk) 22:52, 22 August 2022 (UTC)

@JTanner (WMF): I just wanted to say that I'll be busy off-and-on for a while but that I'm still checking my email and that it's nice to see the updates on the phabricator tickets as they come in. It's nice to know that my feedback was appreciated. Clovermoss (talk) 21:59, 31 August 2022 (UTC)

@Clovermoss:I'm really happy to see this! Take all the time you need, if you notice anything along the way please don't hesitate to reach out and I'll continue to tag you on things that may be of interest. JTanner (WMF) (talk) 13:24, 6 September 2022 (UTC)

@JTanner (WMF): I commented on one of the phabricator tickets recently [10]. I think I did that right and that your team should be able to see it? If I didn't, please let me know what I should be doing differently. I'm not sure if this technically counts as a seperate issue or if the fix would be the same, but it's similar to the other problem. Essentially, you have the "thank" button for bots and then you get the error message if you try to do so. Clovermoss (talk) 13:26, 6 September 2022 (UTC)

@Clovermoss great point! Its in my review column so that was perfect timing, I'm testing those tasks now. I'm going to kick the task back or make a separate task and put it on the board in ready for development for it to be addressed based on the discussion we have in our planning meeting today. Also, if you download the Beta version of the app and have a moment to test out the editor to see if the speed/words jumbling issues improved that would be really great! JTanner (WMF) (talk) 13:34, 6 September 2022 (UTC)

@JTanner (WMF): I'll try to do so sometime this week. Just to clarify, by beta do you mean a seperate app that I should download or is this something I can change/opt-in through settings somewhere? Clovermoss (talk) 13:38, 6 September 2022 (UTC)

@Clovermoss great! if you want to wait it should be in the production version (the version of the app you have) the end of this week. The Beta version is a separate app to download from the same app store you got the production version of the Wikipedia Android app. It is called Wikipedia Beta. It is our test version of the app that anyone can download if they want to test out features and improvements early with the understanding that sometimes there may be bugs because it is a test space for us. JTanner (WMF) (talk) 13:43, 6 September 2022 (UTC)

Yet another break in this massive discussion

@JTanner (WMF): So I downloaded the beta version of the app but I haven't done anything with it yet. Right now I have both the normal app and the beta one... will that cause issues going forward? I'll try to test what you suggested sometime this weekend. I also think it'd be a good idea to compare the two versions of the app if I'm able to.

I still get the option to thank IPs in diffs even though the phabricator ticket has been marked as resolved? Is that just because whatever your team did to fix the issue hasn't been released yet or because something is going wrong on my end? Clovermoss (talk) 11:49, 8 September 2022 (UTC)

Hey there @Clovermoss, having both versions of the app shouldn't cause any issues. In the Beta version of the app you shouldn't have an option to thank an IP editors anymore, that update to the production version of the app should've went through a few hours ago. Have you updated the production version of the app in the last 12 hours? Also test it out in Beta because for sure the change is there. Thank you for dedicating so much time to giving feedback and testing things out. It is really appreciated and helpful! JTanner (WMF) (talk) 19:39, 8 September 2022 (UTC)

@JTanner (WMF): So I made sure I had downloaded the latest update from the app store for the production version and everything seems to be fine so far. I was able to type this reply at a normal speed and not letter by letter, so that's a good sign. I'll let you know if anything changes. Something else I noticed lately is that the "more" option to edit a page in its entirety doesn't show up until I click a section, at least on user talk pages. I'm not sure if maybe this is something that should be showing up as an option but isn't? I see some options for both (with or without clicking a specific section), just not for this. Thoughts? I also noticed that the section below this one kept changing size, that didn't happen before. I can email you a video of what exactly is happening there if you think that would be useful. Clovermoss (talk) 03:12, 11 September 2022 (UTC)

Hey there @Clovermoss, how are you doing?!

Apologies for the gap in time, I had some life changes where my hours were a little constrained.

• "the "more" option to edit a page in its entirety doesn't show up until I click a section" - So we didn't make topic level of the talk pages editable, as you correctly pointed out, when you choose a topic on a talk page you can edit the source. Does this feel rather limiting? I ask so that we can create a feature request task to capture this.
• "I also noticed that the section below this one kept changing size, that didn't happen before. I can email you a video of what exactly is happening there if you think that would be useful" - Yes it would be helpful to have a video to see the change you're experiencing so that we can properly replicate and diagnose what is happening there

JTanner (WMF) (talk) 15:23, 27 September 2022 (UTC)

The option to thank IPs in diffs has also gone away. So the update seemed to fix that. Thanks! Clovermoss (talk) 16:10, 11 September 2022 (UTC)

@JTanner (WMF): I've noticed some other issues lately. The overlapping text issue is worse and not just appearing in lists or maybe I'm just noticing it more. If I click my edits and view them, I see the ones from Commons but get an error message about the page not existing if I click on it. Which makes sense because this is the Wikipedia app but maybe the error message could be changed or only local contributions could be shown? This feature might work better if someone regularly contributes to different language wikipedias, though, since you can switch between language versions while browsing? Another thing is that in regards to my watchlist, the default setting is to show changes made by me, too. I'm not sure if there's a way to change this like in my settings or something because on desktop edits made by yourself are automatically hidden and I'd prefer that to be the default option. Clovermoss (talk) 20:56, 12 September 2022 (UTC)

@Clovermoss

• "The overlapping text issue is worse" - Is what you are experiencing, you start typing in the editor and the letters are not being captured fast enough so it is jumbling? Or is overlapping happening in the review or something else? It would be helpful to see a screenshot for more information
• "I see the ones from Commons but get an error message about the page not existing if I click on it" Are you clicking on the commons contribution from watchlist in the production version of the app? I am going to come back with a ticket, because I see other issues related to contributions made not showing up
• "watchlist, the default setting is to show changes made by me, too. I'm not sure if there's a way to change this like in my settings or something because on desktop edits made by yourself are automatically hidden and I'd prefer that to be the default option"- I will discuss this one with the team today and based on the conversation I'll make a ticket and share it with you

JTanner (WMF) (talk) 15:54, 27 September 2022 (UTC)

@JTanner (WMF): Another thing I noticed lately is that it's impossible for me to see diffs that other people link. The spot where they would be is a blank space. I notice that the same thing happens with external links. Clovermoss (talk) 20:11, 22 September 2022 (UTC)

@Clovermoss: Tried investigating this a bit and could see diff links when posting it on a talk page. When clicking on the link it did however take me into Mobile Web view which I'll bring up with the team today. I tested in dark mode on the latest production version of the app. Are you seeing a different behavior than this? JTanner (WMF) (talk) 16:01, 27 September 2022 (UTC)

@JTanner (WMF) It's okay. I figured you were just really busy and that's why it was taking awhile. As for diffs, I don't see them on talk pages at all. It's literally a blank space where they should be. I can take some side-by-side screenshots if that would be helpful (one in edit mode where you can see the link to the diff and then when you're not able to see it at all). I'll try to do that video within the next few days. I've been really busy this week trying to catch up on things I should've done last week. Clovermoss (talk) 20:31, 27 September 2022 (UTC)

@JTanner (WMF): Another thing I've noticed lately is that every time I try to edit a page like AfDs, RfDs, MfD, etc, it causes the app to crash. I'm assuming you'll want a video of that too when I can get around to it? To your question about the production version of the app, yes that's what I'm referring to. I still have the beta version downloaded but I haven't really done much with it yet. Clovermoss (talk) 01:13, 28 September 2022 (UTC)

@JTanner (WMF): I've sent an email showing the AfDs causing the app to crash issue, I still need to get around to a video example of the other issue. I've also noticed that the app in general has been crashing more lately, but with AfDs and other XfD-related venues it's an always thing. It happens the second you alter anything, you can view and even scroll in edit mode relatively fine before that point. Clovermoss (talk) 00:45, 29 September 2022 (UTC)

@JTanner (WMF): There seems to have been an update recently. The general experience of the app is better, I like it :) There's still some things that could do with improvement, but I just wanted to say that I appreciate your dedication to making things better. Did you recieve the video I sent to you? It's okay if you're still investigating the issue, I just wanted to make sure you actually have access to it. Something else I noticed lately is that diffs differ between IPs/registered users. For a registered user, you have the thank button and then undo. For an IP, you have a rollback button and that's it. I'm not sure how this would appear to someone without the rollback permission, but it makes me a bit uneasy because there isn't a visual difference between the two and they're in the relatively same spot. You can still undo IP edits if you click the three buttons at the top and go through the options there but I just think it'd be better if the undo formatting was more consistent regardless of who made the edit. Your thoughts? — Preceding unsigned comment added by Clovermoss (talk • contribs)

Late 2022 and early 2023

Hi JTanner (WMF), it's been awhile since we talked. I recently noticed that my most recent comment wasn't signed so you likely didn't get the ping/notification about my comment here. Did you get a chance to see the video I sent you of the app crashing whenever I tried to participate in an AfD? The issue seems to be resolved with the latest update so that's good at least. I admit I'm curious about the why, though. Do you think you might be able to explain it me? I think I'd find it fascinating. Anyways, how have you been? Anything exciting your team has been working on? Clovermoss🍀 (talk) 19:40, 5 November 2022 (UTC)

@Clovermoss!! Hi how are you?! I hope you are well. I am doing well, catching up on some things, I was out then a bunch of folks on the team have been in and out. I think with it getting cold in some parts of the world the weather is getting to some of us. I did miss the ping! So thanks for pinging me again. We've made some updates to the editor recently including improving the syntax highlighting, making adding an image from Commons easier, and added the option to copy the link to a diff in a way where even if it doesn't appear in the bottom sheet, it is still accessible. You can see all of the tasks via this link. The engineers are doing some less visible changes right now to improve overall performance while we research what anti-vandalism tools we could bring into the apps. We are learning from tools that were created by volunteers and discussions about the pros and cons of those tools. If you have ideas related to anti-vandalism definitely feel free to contribute on the task, soon we will create a MediaWiki project page about it. Were you able to make the video about the diffs/are you still experiencing that challenge? Thank you for sending the video over about the AfD bug, our lead engineer @DBrant (WMF) actually fixed it, he's going to share more details of how he was able to fix the bug.

Regarding the diff view and rollback this is really helpful to know, let me share the rationale and tell me what you think. Since we know that IPs can't be thanked, the logic of the code hides the thanks button for IP edits. We felt it was deceiving to show a Thanks button for IP edits only for it to fail, especially after your helpful observation. For someone who who does not have rollback permissions they'll see Thanks and Undo for Registered Users Edits and for IP edits they'll just see undo. We tried putting Rollback, Undo and Thanks all in one are of the screen but it was really crowded and we wanted to be mindful of phones that have small screens so we hid undo in the overflow menu for folks that have rollback permissions. As someone that uses the app and have rollback permissions, would you have preferred seeing undo and clicking the overflow menu for rollback or a different variation? I can imagine the configuration moving being confusing.

But I am thrilled to hear that you feel the app is better overall, and you like it, this news has made my week! If there are other things you've noticed or do notice please keep sending it our way so we can continue to improve the app. When our Designer is back on Android we are planning to make some updates to make sure we are keeping up with the requirements from the Google Play Store but also revisit some of the things you surfaced earlier.

Look forward to hearing how you're doing! JTanner (WMF) (talk) 19:51, 18 November 2022 (UTC)

Hey there @Clovermoss, nice to meet you.

That crash didn't really have to do with participating in an AfD discussion, but rather with the native wikitext editing interface itself, regardless of which article is edited. The crash occurred when putting the cursor all the way at the bottom of the text and pressing Enter.

When building our native app, we care about squeezing every bit of performance from the Android platform, so that features like editing could be more responsive and usable than the web interface. This often involves some complex logic and system-level optimizations. And although we do as much testing as we can, the occasional crash can sometimes slip through. If you're curious, this is the exact change that fixed the crash. Interestingly, it was a single incorrect character of code that was causing the crash, not even a single line. DBrant (WMF) (talk) 20:09, 18 November 2022 (UTC)

Nice to meet you as well DBrant (WMF). Thanks for the explanation. I'm trying to learn more about how stuff like this even works and it's surprising to find out that one character being wrong can cause this kind of problem. I didn't even notice the scale of it but it makes sense looking back that I only noticed the issue there. I don't mess around with stuff like categories that much, I'd typically use the reply button on talk pages, etc. I'm glad it's fixed now, though. Thanks for following up on that. Clovermoss🍀 (talk) 08:29, 19 November 2022 (UTC)

@JTanner (WMF): It's nice to hear from you again as well. Honestly it's no surprise you didn't get my ping because I forgot to sign my comment. Not something that happens very often but it is something that causes pings to fail (see Help:Notifications). The reply tool must have spoiled my habit of remembering the four tildes. Anyways, it really is nice to hear from you again, I'm seriously ecstatic each and every time.

In regards to the undo/rollback/thank setup, I don't think putting rollback on the overflow menu is nessecarily the best idea because usually the advantage of rollback compared to a regular undo is the speed of it. On the other hand, people worry about accidently using the tool when they don't want to. People will sometimes customize their watchlist on desktop to get rid of rollback for this reason (see Wikipedia:Customizing watchlists#Remove or modify the rollback link). That was my main concern with the variance in the button placement on whether you're reverting an IP or another registered user as it's possible that mistakes are more likely to happen that way. I understand your point about how having everything together would be crowded, though. Maybe some way of customizing it in settings would be helpful? Or making the buttons vertical instead of horizontal? Something like 'undo' and 'rollback' for IPs and 'thanks', 'undo' and 'rollback' for registered users? That way, if someone accidently presses something, it's more likely to be undo instead of rollback.

I have no idea how realistic my ideas might be though. I'll think a bit more about how countervandalism might work better on the app. Personally I have a hard time seeing how it would work well without some way of being able to view recent changes like you can on the regular mobile version of the wesbite or on desktop view. I'm assuming that there's a reason this hasn't been done on the app already yet, likely it would cause some sort of performance issue? Clovermoss🍀 (talk) 10:02, 19 November 2022 (UTC)

@JTanner (WMF): I've also emailed you the video about the overlapping text issue I mentioned to you before. I kept forgetting to actually demonstrate it but your comment reminded me that was something I needed to do. Hopefully it's helpful to understanding what is going on. I used my talk page as an example because I noticed that it's more likely to happen on pages with substantial content but it also happens when you're editing an article too. But I think the video has a few good examples of what it can look like and how line breaks that involve weird cursor movement can kind of fix it. In the future, would you prefer it if I emailed any videos I notice of technical issues to DBrant (WMF)? I emailed you because that's what I did the last time but if it'd save some time I can email him directly. Clovermoss🍀 (talk) 14:50, 19 November 2022 (UTC)

You're very welcome to direct any technical issues (regarding the app) to me. DBrant (WMF) (talk) 02:49, 22 November 2022 (UTC)

@DBrant (WMF) Thanks. I will wait to see what JTanner (WMF) has to say in regards to her opinion on what's best as well, but in the meantime, I have pinged you to an app-related technical issue at the Teahouse. Clovermoss🍀 (talk) 03:13, 22 November 2022 (UTC)

JTanner (WMF) By the way, I noticed that the 'views' button isn't statically zero at all times anymore so that seems to be working now, too. I also wanted to say that I'm open to doing something like a video call like what's mentioned here [11], only in regards to mobile editing and all that. It's just that there's a few things that I have found difficult to express in my essay and in a way it's easier to just kind of demonstrate what I mean? I feel like I'd be more useful with some of my observations that way. If you're not comfortable with that, I'd understand, I'm just throwing it out there as an option. I promise to be polite and courteous, if that helps. Clovermoss🍀 (talk) 05:41, 3 December 2022 (UTC)

DBrant (WMF) and JTanner (WMF); I hope you've both had a good holiday season. Heads up, a version of my essay is going to be published in the December edition of the Signpost. Current version is here: Wikipedia:Wikipedia Signpost/2023-01-01/Essay. If there's anything you want to say, let me know. I hope you feel like I've been fair. I really do appreciate that you've listened to my feedback and I hope this doesn't get in the way of that. I'm still planning on using the app and seeing how it gets better over time :) I even have a new phone (haven't used it yet but it might be interesting to see potential differences). Clovermoss🍀 (talk) 05:54, 26 December 2022 (UTC)

Hey there @Clovermoss how are you??!! I returned this week after being out for a few weeks. I was able to get some things with my health under control, and wow what a wonderful way to jump back in.

Your Signpost is outstanding. Thank you for writing such an insightful recap of our interactions and sparking really great conversation. I've had both the Android and iOS teams read it this week and we all found it incredibly helpful and have circulated it with other colleagues that work on Mobile Web. I believe you were more than fair and an amazing contributor to our team (I hope it isn't overreaching to say it feels you are a part of the team). You posting your honest feedback would never impact our collaboration.

Congrats on the new phone, what model is it?

To get back to some of your questions and thoughts here, I'm happy to answer any questions, often the "why does this existing thing work this way" that may feel like a bug and not a feature I lean quite heavily on @DBrant (WMF) if I don't immediately know the answer. He's deep in the code base and one the most expert level engineers at WMF so he'd be able to answer those questions (if it is related to Android) faster at times since he is the one I may go to to ask. For the "why doesn't this feature exist, or can we change or add those features" I'm 100% your person. DBrant also answers the "how hard is it to do this technically" questions and I'm the other half of the pie of "when can it be prioritized" or "is it not prioritized for a specific reason". I hope this all makes sense!

Thanks for this feedback on Thanks vs. Undo vs. Rollback. I made a task (T326894) so I can work with the team designer to rethink the implementation of this. I made you a subscriber so you can see a few proposals of changes and give us feedback as we think it through.

We are exploring some designs around a recent changes view for anti-vandalism but also comparing it to more advanced volunteer created tools like twinkle and huggle. I hope to tag you on those in a few weeks to hear your thoughts. Are you still seeing the overlapping issue by the way?

Its like you read my mind about the video call. Our Community Relations Specialist @ARamadan-WMF is going to post on signpost and propose a few touch points and see what you and others interested in apps prefer. Amongst the choices is a monthly or quarterly video call amongst other options. We can do one or a mix of whatever you all find most interesting but either way, I'd love to have a video call and connect synchronously. JTanner (WMF) (talk) 00:12, 13 January 2023 (UTC)

@JTanner (WMF): It's nice to hear from you again. I hope your health issues have been getting better and the new year has been off to a relatively good start. I got a Pixel 6a. I wanted a better phone camera and it seemed to be leagues above my current phone so that's what I went with. I'm glad my feedback/Signpost article feels fair. I've always appreciated your continued discussions with me and I didn't want that to change.

I've seen the emails for updates on the phabricator tickets and it's always nice to see that your team is working on trying to make things better. As I said earlier, I'm open to a video call, whether it's more of a one-on-one thing or in a group. Clovermoss🍀 (talk) 00:24, 13 January 2023 (UTC)

Hi JTanner (WMF) and DBrant (WMF). So when pages I create show up on my watchlist on the app... I noticed something that's a bit weird. Specifically, when redirects have been reviewed. It doesn't really say what the change actually is in my watchlist but I get a notification about it. But I tend to check the redirects on desktop immediately afterwards because even though there's an invisible diff so-to-speak, there's a + or - bytes. Even if the person who reviewed the page hasn't edited it at all and it's just me. It's really quite bizzare. Here's some examples: Chimeral variegation – my watchlist shows that Ozzie10aaaa added 179 bytes [12], Moss wrack – my watchlist shows that SunDawn removed 15 bytes [13], Hardenburg's burrfish – a removal of 268 bytes [14]? I've only ever observed this issue on the app and it's confusing. I can give more examples if it'd be helpful. I'd also be curious to understand the why behind what is actually going on here, too. Clovermoss🍀 (talk) 21:46, 30 January 2023 (UTC)

Hey there! Hope you're having a great year, and thanks as always for using the app and letting us know your observations. It looks like you've uncovered a genuine bug (I've created a Phab task: phab:T328453). In the case of watchlist events that are not actual "edits", such as page curation logs, there's no concept of a diff, and therefore the number of bytes modified becomes undefined. And yet, our native component that shows watchlist events always assumes that there must be a byte count, so it takes a byte count from an undefined source. At first glance it looks like it's taking the byte count from another random item in the list of watchlist events, which... it should not be doing! Anyway, the task I linked will fix this issue, as well as improve the appearance/behavior of watchlist items that are not plain old edits. DBrant (WMF) (talk) 16:30, 31 January 2023 (UTC)

Thanks DBrant (WMF) for taking a look. Random items from the rest of the watchlist seems like a reasonable explanation for what's going on... I spent a while trying to figure out if maybe there was some pattern and I just couldn't.

JTanner (WMF) Thanks for subscribing me to the relevant phabricator ticket. I really do appreciate every time you do so. I forgot about what you said the last time about how DBrant is more for technical issues and you specialize more on brainstorming/explaining how things work way. Would you prefer it if I didn't ping you on obvious technical issues like this in the future or is it okay to keep pinging you both to guarentee more eyes? I also noticed that you mentioned that someone else was going to do something regarding the Signpost piece. I didn't see them comment there – maybe they reached out to people individually or just haven't got around do doing anything like that yet? Clovermoss🍀 (talk) 19:27, 31 January 2023 (UTC)

Hi JTanner (WMF). I think this would count as a feature request, so I'm pinging you. I think I mentioned this before but I'm not sure if anything happened beyond investigating how pageviews on the watchlist worked? Or perhaps I'm misremembering/misunderstanding what happened before. What I'm getting at is that as someone who alternates between the app and desktop editing, it'd be nice if I could have some way of marking changes on my watchlist as read on the app (or automatically updating this) so when I'm looking at my desktop watchlist it is more relevant to me. Or is this something that I should try adding to the chorewheel that ARamadan-WMF has mentioned? Maybe that would be the more appropriate venue for something like this? Clovermoss🍀 (talk) 12:29, 8 February 2023 (UTC)

Hello @Clovermoss, I created a ticket on Phab. to investigate this issue and added you as a subscriber; here's the link you can follow. ARamadan-WMF (talk) 12:03, 15 February 2023 (UTC)

@DBrant (WMF): I'm not sure if this would fall under phab:T315347 or if it should be its own seperate phabricator ticket, but I noticed that there's another circumstance where redlinks show up as blue links. Specifically, if you're looking at replies on a user talk page. [15] That link was actually blue which worked well for the joke, but it shouldn't have actually happened. Clovermoss🍀 (talk) 08:38, 6 March 2023 (UTC)

Another valid observation, filed at phab:T331299. DBrant (WMF) (talk) 15:14, 6 March 2023 (UTC)

@DBrant (WMF): I found another one, I think! So I get an error code when trying to look at Talk:2023 Hamburg shooting on my watchlist, specifically "Empty list doesn't contain element at index 0". I'm not sure what that means, but I can provide a screenshot of the error if that helps. If I had to guess, it's likely a result of the article being moved a few times. Clovermoss🍀 (talk) 07:58, 11 March 2023 (UTC)

Acknowledged and reproduced! Some of our watchlist and diff handling code seems to be too dependent on page titles, when it should be dependent on page ids. When we attempt to get a diff of a revision that has been moved (based on the page title), it fails, but if we use the page id, it will work. (phab:T332173) DBrant (WMF) (talk) 14:06, 15 March 2023 (UTC)

@DBrant (WMF): The app has been behaving very strangely lately. The best way I could describe it is that it's acting as if I'm logged in and logged out at the same time. I plan to email you a video of what exactly is happening. Could you please look into it? Clovermoss🍀 (talk) 09:19, 15 March 2023 (UTC)

Thanks; I've reviewed the video, but I'm not seeing any unusual behavior on our test devices. It's possible this was a temporary glitch on the server side that was producing errors. Is it still behaving this way now? DBrant (WMF) (talk) 11:47, 15 March 2023 (UTC)

@DBrant (WMF) No, it stopped. I've seen some of the similar weirdness before but never all of those issues together. Do you want me to keep track of the dates and times I experience anything unusual? Clovermoss🍀 (talk) 12:29, 15 March 2023 (UTC)

Sure, any additional details would be helpful, although the thing that helps us the most is steps that are consistently reproducible. In this kind of case the possible causes are too broad to attempt to track down without more data. DBrant (WMF) (talk) 13:56, 15 March 2023 (UTC)

DBrant (WMF) So with this morning, the only thing I did was open the app. That's the only thing I can think of in regards to what might have caused it, which doesn't really narrow anything down. Before today, the only thing weird I really saw happen was my watchlist occasionally telling me that I needed to login (while being logged in). Typically, waiting a few minutes and refreshing a few times would make that go away. The other issues (e.g. saying I only have 400 contributions on the homepage, etc) were completely new. I'll keep you updated if I notice anything that you might find useful. Clovermoss🍀 (talk) 15:26, 15 March 2023 (UTC)

@DBrant (WMF): This morning I noticed that whenever I tried to edit a specific section of something in the Wikipedia namespace on the app, it would prompt me to edit the next section and not the one that I was actually trying to edit. I tried it a few times so I'm certain it's not a misclick on my end. I'm not sure why this happens or if it's a broader issue, but I thought I'd let you know. The page in question is ~300,000 bytes and I've noticed before that sometimes things act weirdly when trying to edit large pages, so maybe that has something to do with it? Clovermoss🍀 (talk) 14:15, 23 March 2023 (UTC)

@JTanner (WMF): I've done more experimenting with this specific technical issue. So far, I've only experienced it at WP:ANI and WP:AN. I've tried other random pages that are huge byte-wise both inside and outside the Wikipedia namespace but haven't had much luck replicating this issue elswhere? When it does happen on a page, it happens consistently. I can keep you updated if I notice it elsewhere as time goes on. Hopefully you'll be able to find out what's causing this? Clovermoss🍀 (talk) 12:09, 15 April 2023 (UTC)

@JTanner (WMF): I've noticed another technical issue. If you try to search for a specific user's userpage and it's a global one, it's impossible to see it. Clovermoss🍀 (talk) 11:34, 26 April 2023 (UTC)

Hi again, and sorry for the delay — Yes indeed, searching within the app will only search your current language wiki, so if a user doesn't have a userpage on that wiki, it won't find anything. (And even if it was able to find a global userpage (presumably on meta wiki) the app wouldn't display it because it's not on a wikipedia.org domain.) DBrant (WMF) (talk) 12:59, 8 June 2023 (UTC)

@DBrant (WMF) Thanks for getting back to me. There was one other thing that came up since the last time we talked if you scroll up a bit. Your thoughts on that issue would be appreciated. I was also wondering if you're aware of when the next office hours for the app are? I'd like to attend again whenever they're happening. Clovermoss🍀 (talk) 19:17, 8 June 2023 (UTC)

Yes, we're aware of that issue as well, but the cause of the problem is way upstream from the app. The short answer is that certain pages (outside the Main namespace) that transclude a lot of various templates seem to throw off our server-side parsing logic, but we're working on it. And really, the proper solution will be to treat pages like WP:ANI as a "talk page" and present it in our native Talk interface, as we do with user talk pages. As for the next office hours, I'll defer to @JTanner (WMF) or @ARamadan-WMF. DBrant (WMF) (talk) 20:39, 8 June 2023 (UTC)

@DBrant (WMF) Thanks for all you and everybody on the team does. I know you mostly hear from me when I'm complaining about something but my experience with the app has gotten drastically better over time and I recognize that I'm likely an outlier when it comes to a lot of the stuff I try to do with it. So I just wanted to reiterate that I appreciate my feedback actually being listened to.

The only other potential issue that immediately comes to mind right now is the view counter that shows up on the homepage. It very rarely displays a number other than 0. Maybe something is going on there? Clovermoss🍀 (talk) 01:30, 9 June 2023 (UTC)

Hello, @Clovermoss,

I will ensure that you receive the necessary information regarding our upcoming office hour scheduled for September. ARamadan-WMF (talk) 11:58, 12 June 2023 (UTC)

Mid to late 2023

Hi DBrant (WMF). I just wanted to say that I noticed barnstars are actually somewhat visible through the app now and I appreciate the change. It's always cool to see progress as it happens! Clovermoss🍀 (talk) 20:06, 3 July 2023 (UTC)

@DBrant (WMF): I also wanted to say I appreciate the feature of being able to search my watchlist. Maybe it's a setting I haven't been able to see elsewhere that does exist but it seems unique to the app. It's really cool. Clovermoss🍀 (talk) 16:57, 21 July 2023 (UTC)

DBrant (WMF) I just emailed you a video of another technical issue I noticed. I don't know the best way to explain it but using bold text can cause the app to act weird in a relatively predictable way and it doesn't stop unless you just get rid of the bold text. Hopefully you can figure out what's going on there. I hope you've been doing well lately. Clovermoss🍀 (talk) 12:58, 6 August 2023 (UTC)

@DBrant (WMF): If you were able to figure out what was going on after I sent the video, would you be able to explain it to me? I'm curious even if I don't always completely understand the why. Clovermoss🍀 (talk) 02:54, 15 August 2023 (UTC)

When we perform syntax highlighting, we only parse the current viewport of the wikitext you're looking at. This is for performance purposes -- if we were to perform real-time syntax highlighting on the entire wikitext, especially if it's a larger article, it would become sluggish. The downside of this approach is that there can sometimes be edge cases where the end of a syntax block appears near the top of the viewport, but we parse it as the beginning of a syntax block. If you look closely at the top of your video, there is a triple-quote ''' which is really the end of a bold block from further up, but we're assuming it's the beginning of a bold block, so when you type another triple-quote further down, the whole thing gets parsed as bold. DBrant (WMF) (talk) 15:23, 15 August 2023 (UTC)

@DBrant (WMF): Thanks for explaining that. :) I was also wondering why the pageviews on the homepage almost always says 0? I remember one occassion months ago where it briefly said otherwise but 99% of the time it's stuck at zero which doesn't seem right when I'm editing highly visible pages. It's a bit of a confusing metric. Clovermoss🍀 (talk) 18:55, 15 August 2023 (UTC)

Agreed that the pageviews number is confusing, and the whole "Edits" stats screen is on our queue to be overhauled at some point. The number of pageviews actually refers only to articles whose Wikidata description you've edited. This is a holdover from the early days of the Edits screen, when it was called the Suggested Edits screen, and when the only type of suggested edit was Wikidata descriptions. DBrant (WMF) (talk) 19:56, 15 August 2023 (UTC)

@DBrant (WMF): Thanks for clearing that up. Yeah, I agree that's a good thing to overhaul at some point. While we're both online at about the same time, I was wondering if anything has been planned for the September meeting? Like a date being set? It's mid-August and I haven't heard anything yet so I just want to make sure I haven't missed anything. I expect my September to be fairly busy so a bit of a heads up ahead of time would be useful. Clovermoss🍀 (talk) 20:01, 15 August 2023 (UTC)

Hey DBrant (WMF). I found another issue that comes from a notification when I get thanked for editing a redirect page. I'm about to email a video of that to you too. Hopefully it helps. Clovermoss🍀 (talk) 00:04, 20 August 2023 (UTC)

A long-standing task: phab:T68480 DBrant (WMF) (talk) 11:36, 20 August 2023 (UTC)

@DBrant (WMF): Thanks for letting me know. It's good that I'm not always the only one to notice these things. Clovermoss🍀 (talk) 11:37, 20 August 2023 (UTC)

@ARamadan-WMF: Remember how I said I wanted to show you and Jazmin something that may be useful in regards to the upcoming antivandalism feature at the office hours? There was a particular situation I was thinking of in that moment but I incorrectly remembered what template I was having issues with at the time [16][17]. In that one, the substituted template creates its own heading and since you can't insert a template on the app without a subject line on a redlinked page (because you're forced to use the new thread feature) you end up with two. This isn't a problem if a talk page is already created (as it's possible to just edit the entire page). Anyways, I am going to send that email soon showing a recording of something else that might be of use to the app's team. I kind of showed the switch from the app to desktop mode on browser that happens when I revert someone. It's somewhat convoluted and might be helpful in implementing that new antivandalism feature you're working on. Implementing Wikipedia:User warnings as some sort of option in the saved messages thing would probably be what the average Wikipedian would look for first, at least on enwiki. While I definitely write personal messages for people sometimes (especially if the issue is more complex), by definition, actual vandalism tends to be fairly straightforward to deal with. More personalized messages are more useful when someone is editing in good faith and not merely writing stuff like this. Hopefully my feedback here is useful. Clovermoss🍀 (talk) 02:49, 28 October 2023 (UTC)

Thank you, @Clovermoss, this has been definitely helpful; I created this ticket T350583 with what you mentioned, and I added you as a subscriber. ARamadan-WMF (talk) 14:19, 6 November 2023 (UTC)

@ARamadan-WMF: I'm glad that this was helpful feedback. I noticed something else just now that might be useful to know about. There isn't a "watch" option for templates while you're on the app. Sometimes templates have changes (e.g. Template:Centralized discussion) so it would be useful to watch pages like that without having to switch to desktop. Thanks, Clovermoss🍀 (talk) 17:44, 10 November 2023 (UTC)

Please check our software engineer reply to you, here on Pah.

And the second feature suggestion is now in a ticket and will be triagged during the next week.

You're always welcome to share your thoughts and ideas! ARamadan-WMF (talk) 13:05, 22 November 2023 (UTC)

@ARamadan-WMF: Thanks for letting me know! I didn't really have wifi for most of the past week so I'm somewhat behind on things. Clovermoss🍀 (talk) 00:45, 24 November 2023 (UTC)

Hello, again!

You can find the update of your second suggestion here, in the ticket on Phabricator. ARamadan-WMF (talk) 18:27, 28 November 2023 (UTC)

@ARamadan-WMF: I noticed that. JTanner (WMF) subscribes me to phabricator tickets so I get emails about them when they're updated, even if I don't always comment directly on them. Thank you for going out of your way to let me know though, I appreciate the sentiment behind the action. :) Clovermoss🍀 (talk) 18:30, 28 November 2023 (UTC)

@ARamadan-WMF: I'm about to email you another issue I've identified. I'm not quite sure how to explain it, but whenever I try to edit a page that looks like that, it gives me an error. I have to go into the overflow menu to edit the page at all. Otherwise the reply function works normally on talk pages and the like. Clovermoss🍀 (talk) 03:55, 17 December 2023 (UTC)

Yes, I received it and created a ticket for this issue as explained.

You will get notified when it gets triaged by the Android software engineers. ARamadan-WMF (talk) 10:20, 17 December 2023 (UTC)

@ARamadan-WMF: Okay, thanks! I think the next time I find anything I'll try to report it on phab myself, I honestly should have learned how to do that already. It's always a good idea to try and learn new things. Clovermoss🍀 (talk) 11:06, 17 December 2023 (UTC)

It's okay, I am always here :) And I can help you get familiar with Phabricator, too.

Have a good day! ARamadan-WMF (talk) 11:56, 17 December 2023 (UTC)