Cascadia Wikimedians/2018 Wikimedia Conference report

This is a report on the Wikimedia Conference 2018 in Berlin, Germany. The Core Conference took place April 20–22, 2018, with Learning Days for 2 days before that. This report is mainly written as a report back to the Cascadia Wikimedians, but it is intended as a public document.

Peaceray

Joe Mabel giving his lightning talk

Attendees

Joe Mabel (User:Jmabel) and Ray Leonard (User:Peaceray) attended the Wikimedia Conference 2018 as representatives of Cascadia Wikimedians. We also attended the Learning Days before the main conference. Sage Ross (User:Ragesoss), also a member of Cascadia Wikimedians, attending independently, in his capacity with WikiEd, and has written his own remarks in that capacity here. If other members were there in some capacity, please feel free to add yourself here.

Personal note from Joe Mabel

Where my comments here are opinion, rather than fact, I am signing them. I would request that others do the same. Some of my remarks might be seen as criticisms; where I think that is liable to be the case, I have attempted to omit names or other identifying information. If anyone in or out of Cascadia Wikimedians feels feels they are identifiable in a context where they would wish not to be, please let me know and I will remove the reference, and (if desired) request that a Meta admin remove the relevant portion of the history of this page. - Jmabel (talk)

Overview

Group photo of the attendees

Of the two major annual international Wikimedia gatherings, the Wikimedia Conference is the more institutional in focus. The other, Wikimania, is above all a meeting of Wikimedians; the Wikimedia Conference is above all a meeting of people attending on behalf of affiliate organizations, in our case Cascadia Wikimedians. Peaceray and Joe both attended the Learning Days and conference on full scholarships from the Wikimedia Foundation, for which we would like to convey our thanks; Joe stayed on several more days on his own, and consequently participated in what might be described as the informal penumbra of the conference.

The Learning Days (full schedule here) included sessions on topics such as consensus-building, developing partnerships, facilitation, etc. The main conference (full schedule here) combined one track focused on overall movement strategy with parallel sessions on partnerships, capacity-building, etc. Some typical sessions from the latter were "Exploring privilege through autobiographies", "Introduction to financial planning for your organization", or "Wikipedia and digital equity: digital skills for life and free knowledge for the world". Both parts also included numerous lightning talks. Both parts included several sessions related to the forthcoming plans to use Wikidata heavily as a back end for structured data for Wikimedia Commons, all of which Joe attended.

There were 100 attendees for the Learning Days and an additional 200 for the main conference. Many of the sessions went back and forth between having 20–100 people all in one meeting to smaller breakout groups of anywhere from 3 to 12, and then back to the full session.

Meetups and informal discussions

Wikidata meetup group photo

There were also an enormous number of less formal session such as meetups and an even greater amount of completely informal discussion on Wikimedia-related topics and everything else under the sun. It was not unusual for some people still to be continuing informally until 4 a.m., catching a few hours sleep, and then being in a session at 10 a.m.

• Joe and Peaceray both attended the Wikimedia United States Coalition (WALRUS) meetup
• Peaceray attended the Wikidata meetup on Friday, 4/20
• Peaceray attended the GLAM Lunch Meetup on Saturday, 4/21
• Peaceray attended the Wikimedia and Libraries User Group during Sunday Lunch, 4/22

Strategy track

I'm afraid we have little to say about the strategy track other than that they seem to be proceeding apace toward the Wikimedia 2030 plan. In the opening plenary session, we were told that the strategy track was intended to be all-or-nothing: if you wanted to attend that track, you should plan to attend all of it. Since Peaceray planned to attend all sessions about financial planning and Joe planned to attend all sessions about Structured Data for Commons, we both knew we could not attend the strategy track. Only later did we find out that only a small core actually attended all sessions of the strategy track and that others—indeed, a larger number—dipped in an out. Then in the final plenary, organizers lamented that some of us didn't attend the strategy track at all.

I (Jmabel) wish that expectations had been set more accurately up front: I would gladly have attended about half of the strategy track. There were several time slots where the strategy track session looked like the most compelling, and where I attended something of less interest to me precisely because I was trying to conform to what I had been told were the guidelines. This was one of several instances where those who did what they were told to do were disadvantaged. I wish someone had been better at setting accurate expectations in this and a few other matters. This was my first Wikimedia Conference and Peaceray's second. I gather that those who were on, say, their fifth conference knew to ignore such an instruction, as I will next time (if I have an opportunity to do this again). - Jmabel (talk) 09:20, 25 April 2018 (UTC)

Structured Data for Commons, Wikidata help and documentation

As mentioned above, I (Joe) focused mainly on Structured Data for Commons, which also bled out into Wikidata help and documentation. I attended four formal sessions, plus much informal conversation including a 1-on-1 followup on Tuesday with Jens Ohlig of Wikimedia Deutschland. Unless otherwise noted, the present section should definitely be read as coming from me as an individual.

I had come into this somewhat concerned, because I don't think the Commons community has been as well engaged in Structured Data for Commons as we should be. A lot of on-wiki discussion has been derailed by people with fringe views (e.g. over-concern with forma set-theoretic consistency at the expense of user experience, people who are scared by any change at all, etc.). It was very good to get in a room with the people who are actually doing the work.

To my pleasant discovery and slight surprise, the Structured Data for Commons project really looks like they are doing it right in terms of rolling this out in small pieces, trying to start with the best-understood parts first, and in how they are thinking about it generally. For example:

• They are starting with just multilingual captions for photos stored in Wikidata. Also high on the agenda is improving search to be able to carry some semantics.
• No existing Commons file page will be required to convert to Wikidata at all unless the Commons side decides to do so.
• A more tag-like "properties" system will be designed to sit side-by-side with the existing category-based system. While the two are liable to be integrated down the line, and some on the team are clearly hopeful that categories will eventually fade away, there is no intention to impose this new approach as superseding categories.

I and other "Commoners" were able in the sessions to raise a few concerns that I think will be addressed.

• One of the properties that they plan on having is a "depicts" property. Several of us were concerned that there is an important distinction between, say, a photo of a stop sign and a photo that happens to have a stop sign somewhere in it. The team hadn't thought of the issue, but they agree now that it's raised. Two possibilities: some sort of way of grading relevance of a "depicts" property or the ability to indicate what portion of the photo (or what time-section of an audio or video) the property is relevant to.
  • One caution: it may get a little dodgy when one of the depicted element in a photo is something we are allowed to depict only on a de minimis basis. Is it OK to call out that depiction in metadata? Certainly if it is we will at least need some sort of indication that this depicted item is present only on a de minimis basis.
• A lot of the hard-core "Commoners" (like me) still use the text-based commons:Special:Upload instead of a Wizard or an external upload tool. There was no previous plan to accommodate that. We have now worked out at least in principle this can be accommodated via serialization/deserialization.

Discussions in this area also led to discussions that Wikidata is still pretty opaque. John Cummings, Wikimedian in Residence at UNESCO, is doing a lot to address that for organizations with large existing datasets to upload, but that is only one of perhaps a dozen distinct classes of user. I met Tuesday after the conference with Jens Ohlig; sadly, Wikimedia Deutcshland, who are spearheading Wikidata, currently lack a budget for documentation. He agrees with me that that may need to be more on the list in the next fiscal year's budget. I (Joe) intend to put in 20-40 hours on a volunteer basis helping with some aspects of this, but basically I don't do large volumes of volunteer work in areas close to my professional expertise. This will need more like 1000-2000 hours all told, and that should either be paid work or work for someone for whom it would constitute either an avocation or an apprenticeship.

Something I don't think has necessarily been adequately addressed: there may be some performance issues in areas that are now quite efficient. We really didn't get into that, beyond a mention.

Another stray Wikidata idea that went by: the text of a SPARQL query, or a link to that text, could itself be a property of a Wikidata item. For example, right now we have wikidata:Q1837959 (List of Seattle landmarks). In theory, if the Wikidata content is organized correctly, we could have a reusable query to generate such a list in a usable form, and store that as a property of that Wikidata item.

I could write on this area at 10 times this length, but I think this is more than enough for a general report.

Organizational tracks

I, Peaceray, attended many sessions on organization matters.

During the Learning Days, I attended:

• An Introduction to Organizing in the Wikimedia Movement
• Communications Plans
• Financial planning for your organization
• Evaluation Plans and Making it Count
• Logic Models, Program Evaluation, and Strategy

Of these five, I found the Evaluation Plans and Making it Count session to be the most valuable. In particular, the chart on Program Data was very pertinent, revealing an 18–19% retention rate for new editors after several months as a result of writing contests, versus edit-a-thons and editor training workshops that both averaged around 5% retention over the same time period.

During the Wikimedia Conference itself, I attended:

• Community Capacity Map workshop
• How to grow an editing community? 4 different approaches from Wikimedia organizations

Tools and Workflows

Peaceray attended a few learning day presentations that concentrated on tools and workflows.

• Tools Demonstrations

• Programs & Events Dashboard

This tool manages and tracks Wikimedia programs such as edit-a-thons and university and college classes. We have used this tool for the Art+Feminism edit-a-thons at the Jacob Lawrence Gallery at the University of Washington.

(Peaceray then left this session to give his lightning talk.)

• Program and Event Practices

There were a number of small group presentations within the conference room. I, Peaceray, was able to attend three sessions on tools:

• GLAM Data Tools (Alex Stinson)

This covered Data Upload Tools and Batch Evaluation Tools (old version of document is Category and Page Group Tools).

• Developing WikiData and Commons Partnerships (Sandra Fauconnier)
• The Wikipedia Library Card Platform

"The Wikipedia Library provides free access to research materials to improve your ability to contribute content to Wikimedia projects." In the past Peaceray used this to access materials on Questia and HighBeam.

• Wikidata - Commons GLAM Workflows

This workshop addressed the GLAM workflows for data and media partnerships with Wikidata and with (Structured Data on) Wikimedia Commons. It also provided more in depth discussions of the Data Upload Tools and Batch Evaluation Tools previously mentioned above.

Peaceray attended another couple of tools sessions during the conference itself:

• Technical Tools For Programs, Events, and Grants
• Training Modules, Program Monitoring, and Data Collection with P & E Dashboard

The latter session was a useful in-depth demonstration of the Programs & Events Dashboard mentioned above and presented by Sage Ross.

Knowledge equity and diversity

One of the through-themes of the conference was what is being called "knowledge equity". Along with "Knowledge as a service" this is one of the two driving themes of Wikimedia 2030. This somewhat opinionated and personal section should be understood as coming from me (User:Jmabel) individually; if someone else has something to add, please sign.

I find "Knowledge as a service" a bit of a vacuous slogan, because we simply do not have enough agreement across the community as to what it means. "Knowledge equity" is clearer, because even if "knowledge" remains vague, "equity" is easier to explain. We want as much as possible to minimize barriers caused by language, geography, level of formal education, degree of physical ability, etc., both in terms of people contributing to our projects and accessing them in some manner as consumers of content. Relative to "Knowledge as a service" this is less problematic, because "equity" is a much stronger and clearer concept that "service." While we have to wrestle with how to incorporate, for example, communicating knowledge held by non-literate people, and just how we can make as much content as possible available to them, the important thing to me is that we have made the commitment that we should incorporate such knowledge. This means, for example, a firm commitment to creating means by which people who are never going to be able to deal with the apparatus of quasi-academic citation can still make substantive contributions.

Two diversity issues that struck me in a negative way about the conference:

1. Although international representation was generally excellent, and women, while fewer than men, were considerably less under-represented in the Wikimedia movement as a whole, there was a complete absence of people with visible handicaps. In a group of 300, that does not happen randomly or by self-selection. Not one deaf person, not one blind person, not one person in a wheelchair or (as far as I noticed) even using a walker. That means that at some level, in the multiple levels of selection that lead someone to the conference, they are being weeded out. I don't know why this happens, so I have no proposed solution, but I think this needs to be looked into. [yeah, passive voice, I have no clear idea who should take this on.]
   • Aside: the under-representation of women seems to be largely about getting them into Wikimedia projects at all and keeping them there. When it comes to who hold official board positions, paid positions, etc., the proportion of women is considerably higher than the movement as a whole and of course WMF has a female executive director. For women, we don't so much have a glass ceiling as a rickety floor.
2. The formal parts of the conference were far more exclusively in English than I would have expected. Even in sessions with breakout groups, I'm aware of only one session where even one breakout group aligned along linguistic lines (Spanish-speakers at the opening plenary). No problem for me, of course, as a native English speaker, but I was genuinely surprised to see less of this at an international conference than I sometimes see at events in Seattle, where over 85% of the people are native English speakers, as against less than half here. Of course, this was all far more diverse in the informal after-hours discussions.

Joe's lightning talk...

My lightning talk overlapped this area. Roughly, I argued that Wikipedia's academic/quasi-academic approach especially can be very exclusionary, especially with respect to citations, and more so when combined with our "no original research" policy. The often contentious on-wiki environment further exacerbates that. For example, I think here in Cascadia we could get a good number of Native Americans involved in sharing aspects of the history and culture of their tribes/nations that we are currently, sadly missing, but not if it means, as I said, "to argue with a bunch of white people about their own culture."

There is a surprisingly parallel case where individuals rather than groups are excluded. For example, in a music scene where one person has written a memoir, we end up with that person's version of what happened becoming "canonical" because it is easily cited. Even in a biographical piece about someone else, we can cite the memoirist's version and not the article subject's own version. For example, because Mark Arm has written a memoir, and Rob Morgan of the Pudz and later The Squirrels has no interest in doing so unless he can get a good book deal, we end up with an anecdote from Mark about the Pudz and can't even quote Rob saying that Mark has the story wrong.

Fundamentally, to address the contributing side of knowledge equity, I believe Wikimedia needs a way to share knowledge that isn't filtered through academia/books/press. I've jokingly call this WikiAffidavits (first-person statements), WikiDepositions (interviews), and WikiStories (collective contributions created by a defined group, rather than on an "anyone can edit" basis). These contributions might be oral or written. The biggest difference between this and a typical Wikipedia contribution would be that the identity of the author matters. For WikiAffidavits and WikiDepositions it would be a known, named individual with identity verified through OTRS or a similar system. With WikiStories, OTRS would be involved but anonymity could be otherwise preserved; for example, we could verify that someone was an enrolled member of the Makah tribe and is considered an elder, without having to publish that person's name. The resulting materials could be viewed by anyone via Commons or Wikisource, as appropriate; they could also, in turn, be cited in Wikipedia.

... turned out to be less original than he thought

The talk met with a lot of assent, and a few of my ideas actually do appear to be new contributions but, happily, others are already thinking and working along these lines. Whose Knowledge? are already thinking along these lines and have tried some pilot projects, and Felix Nartey in Ghana is working on this with them (sorry, I don't know which of a couple of organizations he's involved with is doing this as an organization) and sounds like he is on the brink of working with numerous tribal groups with well-defined people who can speak for the groups' official views.

I talked at length with Felix and one piece I think we Cascadians could bring to the table is to help with some of the modeling of metadata for this in Wikidata. I've proposed that to the Cascadia Wikimedians on a separate email thread, and will have more to say about that in the future, probably some time in May 2018.

Affiliation and grants

Much of the following may be old news to some of you, but having been far more focused on wiki content than organizational issues, I (Joe) didn't have much clarity before.

WMF affiliates can be categorized variously as:

• Chapters: always regional, always a independent nonprofit organizations, often with an office
• Thematic organizations: equally formal, but not regional. The only one right now is Amical Wikimedia, focused on Catalan language and culture without having to define exactly what geographic area that entails (a very contentious issue, ask me if you want details).
• User groups (like us): usually regional but can be thematic; can have a less formal structure, but still officially recognized by WMF
• Allied organizations without a formal affiliation, such as Wiki Education Foundation (hi, Sage).

Grants have historically been broken down into four categories:

• Rapid grants < US$2,000
• Project grants $2,000–100,000
• Conference grants
• Annual plan grants for affiliates (APGs)

However, during a period of transition, for the next fiscal year there will be no new conference grants (there is, I guess, one exception to this, didn't follow closely) and no new APG grants. This means because we didn't get an APG grant this year, we don't get one next year. If we want to do anything next year that requires money from WMF, it is going to have to be through a rapid grant or project grant.

There was a strong emphasis on reporting results for anything that is done under a grant. To summarize:

• Proposals should have clear goals, concrete outcomes, a good team, and community engagement.
• Reports should make clear "What did you do? Why did it matter? What did you learn" and should include stories.

Along these lines: someone who will remain nameless strongly suggested to me (Joe) that we should try to recruit to our board someone even who is not necessarily a Wikimedian in their own right, but might be interested in administrative/organizational issues.

Lightning talks

#Joe's lightning talk... is covered above.

Among of the best parts of the conference were the lightning talks, especially the ones during the Learning Days. I (Joe) presume someone from WMF will do something comprehensive about this, and I think some of these were even recorded, but a few highlights:

• Andrew Lih talked about doing low-budget conferences: in-kind donations, doing the conference the day before or the day after some other conference in the same city, working jointly with a GLAM and leveraging Wikipedia brand recognition. Also one very specific tip: there is a lot more value in everyone having lunch together than in everyone having dinner together.
• Amir E. Aharoni gave a talk on what he considers the best way to leverage an Edit-a-thon: use it as an opportunity to observe the problems new users have in learning to edit Wikipedia. File phabricator reports on what could be improved.
• Ariel Cetrone talked about engaging new editors and emphasized that "The event starts with your registration experience." Someone new to wikis is likely not to successfully register on-wiki. "Emphasize the familiar... Find your pre-Wikimedia self."
• John Cummings talked about importing datasets, and emphasized the importance of the dataset as such getting a Wikidata Q-code (identity).

Peaceray's lightning talk

I, Peaceray, gave a lightning talk during the Wikimedia Conference 2018 - Learning days #1 - Lightning talks session. Although entitled What Has, and Has not Worked with Partners, it weighed more heavily on what has worked with partners with edit-a-thons. It focused on what a partner can bring to an edit-a-thon, how to use Wikipedia resources and social media to publicize an edit-a-thon (sometimes building on the partner's messaging), working with administrators to resolve IP blocks and account creation, and evaluation tools. You can find the slides on Commons.

Wikipedia Library and partnerships

Jake Orlowitz, en:User:Ocaasi (WMF), gave a talk on The Wikipedia Library, which he founded and heads. There is no point to going into that in detail here—their online documentation is good—beyond saying that if there are journals you need to access for Wikimedia projects and you don't have a ready way to get hold of them, you should definitely look into this: online access on a level close to that you might get from a university library.

Jake also talked at length about forming partnerships, since he's had occasion to form so many. The following notes are a little rough, but I (Joe) hope they should be at least moderately clear:

• ASK -> iterate/refine -> GET.
• Seek mutual benefit (being seen as helping Wikipedia/Wikimedia is often a de facto benefit.
• Try to match the partner's level of formality/professionalism (e.g. you don't want to be the only T-shirt in a roomful of suits, or vice versa).
• Define what is and what is not involved in the partnership.
• Don't talk prematurely to the press. The hardest work is up front, and you don't want to announce something that won't happen.
• Make the agreement
• Do it.
• Report internally first.
• Report/share/brag.
• Iterate/expand.
• Scale/make it a model.
• + difficulty of measuring certain types of benefits

The Partnerships Playbook

Peaceray attended three sessions on establishing, evaluating, and maintaining, improving, or ending partnerships.

• New Beginnings: how to establish a new partnership (The Partnerships Playbook: Chapter I)

In this workshop, we shared our experiences in presenting ourselves as an affiliate, and how we identify and approach new and appropriate partnerships with organizations that share common interests and values and have resources to contribute.

• Putting it on the road: how to set up the daily business of managing partnerships (The Partnerships Playbook: Chapter II)

In this workshop, we shared our experiences on proven communication strategies, monitoring/documentation and reporting, evaluation and input into funding requests, and when and how to communicate what is not working in a partnership.

• Sustainability: Nirvana or Myth when working with partners? (The Partnerships Playbook: Chapter III)

In this workshop, we shared our experiences in ongoing partnerships: How to achieve sustainable and enduring partnerships? How to co-create projects that perpetuate themselves and can make partners independent contributors? How to terminate a partnership when it is not working?

I did come away with some strategies for approaching, establishing, and working with new partners. - User:Peaceray

Miscellany

Katherine Maher samples some of our Theo Chocolate.

• Providing an experienced "buddy" who showed up two days after I did rendered the buddy system useless. - User:Jmabel
• Wikimedia Resource Center is intended as "a single point of entry for Wikimedians all over the world to access specific Wikimedia resources." I think it needs some fleshing out. Some of us may want to be involved.
  • Within that I wanted to emphasize Connect, sort of a way for an individual or group to advertise their availability within the Wikimedia world.
• WMF bans (unlike community bans) are overwhelmingly because of bad behavior, in human terms. The farthest from that which could still result in a WMF ban would be repeated, wilful violation of intellectual property rights.
• The technical side of Wikimedia Deutschland—the one entity in the Wikiverse that is almost as large as the Foundation itself—does not require knowledge of German to work there. If you speak English, that would suffice. All of the tech people have good English, only some have an equivalent or greater level of German.
• When partnering with someone about a data set, work in their interest, not ours, about where the authoritative copy should live. In particular, using the Internet Archive as a repository may give them more control over metadata, deletion/non-deletion, etc. than Wikidata. Of course we'd like to sell them on the advantages of keeping it CC-0, and one of those advantages is that we can copy it into Wikidata, but Wikidata may not be the right place for the "master" data.
• I (Joe) particularly liked the concept that John Cummings and others used of a User Journey rather than just, for example, use cases. The notion that learning certain things might enable learning other things; the emphasis on the path as much as the destination; the notion that any sort of help has to meet the user where the user currently is. - Jmabel (talk) 16:03, 25 April 2018 (UTC)

Looking forward

It looks like WMF are reconsidering whether this conference should continue to exist in this form. One proposal was that the Wikimedia Conference be restricted to governance/strategy, and that the other aspects of this might better be handled by a set of regional conferences.

I (Joe) think that may be a good idea, but I would hope that:

• At least one of those regional conferences is held in conjunction with the governance/strategy conference, so that the governance/strategy people don't become too detached from other aspects of this movement.
• At least once every three years, things are all brought together in one place. The hothouse atmosphere was stimulating and productive, and I don't think it can be reproduced any other way.

Jmabel (talk) 16:03, 25 April 2018 (UTC)

Joe's summary

Another section definitely from me as an individual.

I had an incredible time. I'd highly recommend a conference like this to any serious Wikimedian who can possibly attend. Heck, I'd even recommend that it you can possibly afford it, you do what I did and stay on for a few more days: there were still dozens of us here the day after, and even another day after that many informal discussions still continued.

For me, it wasn't so much the formal sessions as the chance to meet and talk with so many colleagues. A few plenaries are of course necessary, but in general, even within the formal sessions, the ones where we repeatedly broke out to smaller discussion groups, then came back to the group at large were more useful than the "do this exercise" sessions or the "we talk, you listen" sessions. (Still, I want to very much single out one of the latter for praise: "Keeping our enemies closer"—sadly, not much at the link—a great session about finding room to collaborate on some issues even with organizations where we may have conflict on other issues.) And often I got more out of the sessions where discussion groups drifted off topic to solving participants' actual current problems than the ones where a moderator kept pulling us "back on track".

Overwhelmingly, people ranging from individual volunteers to the WMF executive director were incredibly available to discuss almost any issue someone wanted to hash through with them, and were extremely forthcoming about various matters even when that could only be done by going off the record. I had exactly one person who I wanted to talk to who repeatedly cold-shouldered me; for obvious reasons I am not naming that person here.

That said, there is always room for improvement. I mentioned a few things above in #Knowledge equity & diversity (lack of certain diversity), and I alluded to another in #Strategy track (expectations set inaccurately). The strategy track case was not the only case where expectations were set inaccurately, or you were effectively punished if you followed directions. For example, in one session we broke out to ten groups and then were told because time was short, each breakout group should speak a single sentence raising a single issue that concerned them. The first few groups obeyed. Then one went on a little longer. The next one went on longer still. The moderator did not put a lid on it. The next group had two people each speak on several issues. Etc. The earlier groups who did what they were told were effectively almost completely disempowered. I don't want to dwell on that one incident, but it was part of a not-unusual pattern. Basically, I'm fine with a free-for-all and I'm fine with rules, but I'm not fine with unenforced rules that restrict everyone who is trying to cooperate, and do not bind the people who are ready to act like steamrollers.

This last leads me in particular to suggest that there might be more times when there should be a moderator other than the presenter (I believe this was the case for the plenaries and it certainly was for the lightning talks), especially when the presenter is not all that experienced.

So as not to end at all a negative note: the Conference was great, funnest thing I've done so far this calendar year, came away with at least a dozen and maybe twenty useful contacts and a ton of ideas, thanks so much to WMF for the scholarship, and I hope to get to do something like this again.

- Jmabel (talk) 16:03, 25 April 2018 (UTC)