Talk:Wikimedians in Residence Exchange Network/GLAM manifesto 2023

Please feel free to edit!

Please consider the title "manifesto" as sounding more scary or threatening than it should be - it was simply the word we had been talking about for a while, as the goal was: How do we explain the current GLAM wiki state of affairs to someone new to this space, whether it is a new WMF employee, an outsider who doesn't know about cultural or heritage partnerships? The document may look long and polished, but it is not meant to be finished. Be bold, add and edit! - Fuzheado (talk) 16:37, 21 February 2023 (UTC)

New approaches?

Hi @Fuzheado, I've been following some of the conversation and I agree overall with most of the stuff being pointed out (institutions needing dashboards, etc.). So the problem with the "new approaches" section is that it's talking about specific tools and/or problems -- and this has been, to the extent of my knowledge, how things have been tackled so far: focusing on specific problems / tools and solve the most urgent task on the to-do product list. And this is at the core of the problems that we're experiencing right now: investments on the Wikimedia platforms are made with this idea of the "individual contributor" or the "individual editor" and without giving significant attention to the large scale efforts that are conducted inside knowledge institutions to create and contribute to the free knowledge ecosystem (and by knowledge institutions I mean institutions ranging from GLAMs to multilateral organizations like UNESCO and the like).

While I understand the urgent problem that you and others are pointing out about dashboards and such, I think that the most urgent thing that needs to be solved on this front is that we need to understand that large contributions to our multimedia ecosystem are driven by institutions, or if not by them, by power users that need better, more reliable, stronger tools. My impression is that the new approach that we need is to better understand who is driving most of the contributions to Wikimedia Commons (I have some suspicions) and serve their needs better if we want to distinguish ourselves from any other multimedia platforms such as TikTok. Scann (talk) 17:40, 21 February 2023 (UTC)

Definitely put that point of view in the "New approaches" section, as that is most welcome! That section was not meant to just describe small projects, it is to also include big picture concerns that you have rasied. For example, the "Query is the content" point is a very big point saying that the WMF needs to think of "search" not as a support function but as a primary way to think about content in the future. I think that is a good example of a big existential concern that our frame of thinking is too narrow. In that sense, I think what you bring up is also in the same area – we like to think of Wikipedia as "volunteer" which is generally true, but in terms of going to the next level with multimedia content, it has to go beyond this simplistic thinking to address institutional partners and contributions. - Fuzheado (talk) 14:45, 22 February 2023 (UTC)

Possible additions

While the document dwells on what is needed, it also doesn't go into details into what works, and why it matters to GLAM institutions. So how might we add some of these experiences to add to the doc (add more below)?

• Add section on why this matters to GLAM institutions, and what works?
• Add success stories of particular GLAMs and how they are making an impact?
• Add specific user stories?

- Fuzheado (talk) 14:16, 23 February 2023 (UTC)

Off the top of my head, we can talk about some particular cases:

• An AI/machine learning model trained on Met artworks keyword data was used to create a Wikidata game to add depiction info (P180) to Wikidata items for other artworks. (Met)
• Understanding gender and demographics about artists by using Wikidata (Smithsonian SAAM)
• Tate Modern outsources bios to Wikipedia (Tate), and MoMA in NYC using Wikidata identifiers
• Wikidata knowledge graphs as demonstrating cultural connections - Portrait of Madame X (Met)
• Lou Reed archive at NYPL and open access

- Fuzheado (talk) 14:19, 23 February 2023 (UTC)

Raw notes/comments from Telegram

Transcript of ideas from Telegram group GLAM Wiki link, by User:Fuzheado:

I agree fully that we need to try to contextualize GLAM wiki work in 2023. In some ways, we need a "fresh" explanation not only because the leadership of WMF is new to GLAM wiki, but there has also been turnover at many GLAM institutions that need this re-explained. Having a good foundational document would be very valuable.

1. Back in 2012, at the first GLAM wiki boot camp, we also identified the need for this, and created a "one pager" that has been the basis of other such documents: https://commons.wikimedia.org/wiki/File:GLAM_One-Pager.pdf

2. At the same time, in the early days of GLAM wiki, there was tension with the community. At Wikimania 2011, which was the first major Wikimania where GLAM was a focus, we found scribbled notes from attendees who had negative things to say about GLAMs and Wikimedians in Residence, because of the connection with "paid editing." The 2012 "Belfer Center Wikimedian in Residence" at Harvard University was a significant blow to WiR, even if it wasn't technically GLAM. It cast a big shadow over any type of WMF involvement with Wikimedians in Residence. https://www.dailydot.com/debug/wikipedia-paid-editing-scandal-stanton/

One could argue that this one episode created a "radioactive" connection between WMF and supporting Wikimedians in Residence that lasted for years. This one episode also had WMF back off of "earmarked funds" for specific projects from donors, a policy it still adheres to today, probably for good reasons.

Comment by User:Dominic: This was also around the same time as Gibraltarpedia drama.

3. While much of our initial successes and failures involved mainly Wikipedia editing, mass contributions of media files via Commons was much less controversial. The value of uploads could be quite easily demonstrated, as we don't have as many super-productive creators or uploaders of multimedia content, whereas libraries and archives had millions of files ready to donate, many of them easily identified as public domain/expired copyright, and could be used without controversy. However, it took time for these GLAM institutions to embrace a total CC0 policy, which required more work for us to explain the benefits and why they were not going to lose revenue. Key to leading the way was the late Effie Kapsalis of Smithsonian: https://www.slideshare.net/effiekapsalis/sxsw-2016-give-it-away-to-get-rich-open-cultural-heritage

Comment from @scann: I think this paper @sadads3 & @trnstlntk & @Wittylama wrote it's a good overall picture too: https://jlis.fupress.net/index.php/jlis/article/view/95

4. The emergence of Wikidata probably excited the GLAM sector more than anything else. It was not unusual to find librarians and archivists be really excited about Wikidata, when they were only lukewarm about Wikipedia. I recall clearly one librarian stating why: "We don't to interpretation. Wikipedia article writing is not that relevant to us." But Wikidata, descriptive metadata, and authority control is exactly what librarians do, which is why Wikidata has ushered in a new level of engagement with GLAMs that we see at conferences and organizations like LD4, IFLA, et al. We have adjusted pretty well to this new reality, but @scanno makes a good point about formally recognizing institutions as a major contribution partners.

Fuzheado (talk) 14:40, 23 February 2023 (UTC)

Pattypan

Please don't state that Pattypan is end-of-life. Several people use it all the time and several developers are spending time on improving it. 188.149.149.77 13:53, 2 March 2023 (UTC)

Might it be useful then to say that WMF and the original developer considered it "end of life" but that the community continues to help keep it running? - Fuzheado (talk) 14:03, 2 March 2023 (UTC)

Is it true? The only place I have seen such a statement is in OpenRefine's grant application. The current maintainer have also been involved for many years. 98.128.229.28 23:56, 4 March 2023 (UTC)

Search Platform office hours and Telegram March 1, 2023

Here are more notes as a result of the Search Platform Team's office hours, for which @Susannaanas and @Fuzheado were there to discuss GLAM use cases, and WikiCommons Query Service.

• Full Etherpad notes - https://etherpad.wikimedia.org/p/Search_Platform_Office_Hours

Telegram discussion in the GLAM Wiki group starts with this message [1], initiated after the meeting:

Andrew Lih, [3/1/23 12:37 PM]

Thanks for the reminder Susanna - even though we were the only ones there, it was (hopefully) quite valuable to introduce the GLAM Wiki manifesto

Susanna Ånäs, [3/1/23 12:37 PM]

I think it was worth it

Susanna Ånäs, [3/1/23 12:43 PM]

I think we should advocate more for sitting down to plan together with the tech teams. It is quite unknown to me how they get their briefs and assignments – definitely not informed by GLAM work,

Evelin Heidel, [3/1/23 12:43 PM]

and there's little understanding of what the needs are

Evelin Heidel, [3/1/23 12:43 PM]

there's a very vague idea of what GLAM institutions need

Evelin Heidel, [3/1/23 12:44 PM]

or even what they do...

Andrew Lih, [3/1/23 12:44 PM]

Indeed. I pointed the entire set of folks in the meeting to the GLAM Wiki manifesto page to at least have "in writing" the things we have been telling them for a year. They seemed genuinely appreciative that we put it down in a tangible way. I said at least now they can point to a document in their meetings to show what the GLAM Wiki desires are.

Andrew Lih, [3/1/23 12:45 PM]

Again, the document is not complete and more people should add to it! But it's the highest concentration of GLAM wiki needs that I know of so far that is fairly up to date.

Evelin Heidel, [3/1/23 12:47 PM]

yeah I have my opinions on some of these subjects but I don't know if they are relevant. this movement painstakingly documents what we are doing with GLAM institutions, but there's a whole part of the organization that has persistently ignored that as relevant. it sounds like resentment but it's very hard to make your case when it keeps being ignored after bytes and bytes and bytes of documentation

Evelin Heidel, [3/1/23 12:47 PM]

my 2 cents

Susanna Ånäs, [3/1/23 12:49 PM]

Exactly the kind of analysis I heard from the Saami community after the Finnish Government let them down once more with halting a legislation establishing their fundamental human rights.

Andrew Lih, [3/1/23 12:49 PM]

For the record, it's probably useful to relate this nugget from the meeting, not to disparage any one person.

I did ask, "So, are there any stats or metrics they could point to since the launch of WCQS to help understand how good/bad it was doing. Why? The reason for the launch as 'authenticated' was because of scalability concerns and not wanting to have an overwhelmed service like WDQS. So could they provide any insights into whether they are running into issues, or how much "headroom" there is in the system in terms of performance?"

Mike Pham said that since its launch they didn't see any "drop" in users of WCQS, and they were measuring the traffic for 1-3 months afterwards. It's been maybe "6 months" since they've checked (verify this for me Susanna?) but they could check. Mike gave the impression that this was encouraging as the authentication wasn't an obstacle. I pointed out this is a "chicken and egg" thing and that it's not indicative of much at all. Many of us don't even use WCQS or are waiting for it to be unauthenticated, as we cannot use WCQS programmatically and in toolmaking. So the lack of growth and use of it should be a real concern.

Susanna Ånäs, [3/1/23 12:51 PM]

I am sorry I did not know the topic well enough to capture that

Andrew Lih, [3/1/23 12:52 PM]

In that sense, I'm somewhat astonished that they could not cart out a graph right away to show us the traffic and load and stats of WCQS. They did also say we needed to ask the Structured Data on Commons people. So, yeah, that's part of the "Who's responsible for the big picture?" problem

Andrew Lih, [3/1/23 12:54 PM]

https://etherpad.wikimedia.org/p/Search_Platform_Office_Hours

Dominic McDevitt-Parks, [3/1/23 1:29 PM]

The problem isn't just the authentication, but that they did authentication badly. All these months later and they still don't really understand that a tool that provides bulk access to data is primarily for consumption by tools, and not just one-off browser-based queries.

Andrew Lih, [3/1/23 1:30 PM]

It was worrying because it doesn't seem like they have a constant "eye on the prize" of someday (soon) opening it up

Dominic McDevitt-Parks, [3/1/23 1:31 PM]

Like they launched it without even understanding the query service has API access? I still don't understand that part. Because, it's one thing to say authentication is necessary, but then you have to build a way to authenticate if that's youd design... Don't just make it inaccessible, because you require authentication but you don't want to figure out how to let anyone actually do it with their tools. It's the broken design that really gets me, not the authentication itself.

Andrew Lih, [3/1/23 2:09 PM]

If I'm assuming good faith, they were probably working off the simple assumption that people are just interested in coming through the "front door" of https://query.wikidata.org or https://commons-query.wikimedia.org/ - and that's it. What of Listeria, Integraality, Crotos, bots, gadgets, federated search, knowledge graphs, OpenRefine, et al. How do those user stories get recorded and get to the right people prioritizing feature development and launch? That's where we have a big deficit, even though the WMF GLAM/cultural heritage team had done a lot for us in recent years.

Andrew Lih, [3/1/23 2:11 PM]

I'm not even that surprised by this part though. What I am more surprised about is after relating to them what we want/need, and even compelling them to keep the name to "beta," I thought that was enough of a signal that we expected them to monitor and re-evaluate the situation. Actively. But today seemed to indicate there is no such process happening.

Susanna Ånäs, [3/1/23 2:13 PM]

I have heard many positive signals regarding interactions between tech & GLAM, and I am eager for them to materialize.

Andrew Lih, [3/1/23 2:53 PM]

@Susannaanas Just like Python has the concept of "unit tests," we might want to put forth particular end-use scenarios for GLAM wiki that amount to the same. If we cannot pass those unit tests, we should consider it a failure that needs resolution/debugging. One particular case Lydia Pintscher highlighted last year was the importance of being able to "Tweet out a SPARQL query and have it work" by anyone clicking on it, something that fails with WCQS right now.

Susanna Ånäs, [3/1/23 2:58 PM]

Yes, I think there needs to a lot more "professionalization" in regard to partnership / institutional / batch processes. I think it's important that they are seen as core activities, a success factor of the entire ecosystem.

Andrew Lih, [3/1/23 3:05 PM]

I like that idea of "core" and then "secondary" types of concerns that we help define

Fuzheado (talk) 19:44, 6 March 2023 (UTC)

After Wikimania 2023

There were many fruitful conversations during Wikimania 2023, including a panel by WREN to talk about the challenges of the tool chain for Wikimedians in Residence. This document should be expanded to include those. In the meantime, a pointer to the related GLAM Wiki Telegram conversation that occurred after Wikimania 2023 can be found starting at this link: https://t.me/c/1723589369/2842

- Fuzheado (talk) 14:25, 24 August 2023 (UTC)