Talk:Strategy/Wikimedia Foundation/2016/Community consultation/Knowledge/Set 1

Relevant vision document

The I Dream of Content vision proposal is along the lines of the knowledge strategy, and I would like to include it as part of this discussion. --Yurik (talk) 09:59, 13 January 2016 (UTC)

Hello, Yurik. You're certainly welcome to discuss it here on the talk page as the conversation starts on January 18th - both creating your own section through the "button" if you like (and sharing your link that way) and also bringing it up in conversation where appropriate. The visuals especially help make some of your points very persuasively. :) (Just for clarity, while entirely appropriate to link it on the talk page, it wouldn't be appropriate to link to from the main page, since we cannot give equal weight to other voices in that way.) --Maggie Dennis (WMF) (talk) 21:03, 13 January 2016 (UTC)

Slowking4

Response by Slowking4 15:33, 18 January 2016 (UTC)

Slowking4's response to the critical question

one: need to support tools to upload video content

Slowking4's top 2-3 (or share your own idea)

need to support more and better partnerships with GLAM institutions, i.e. fellowships, grants, wiki-GLAM foundation.

Yger

Response by Yger 19:11, 18 January 2016 (UTC)

Yger's response to the critical question

Approach three

BethNaught

Response by BethNaught 21:02, 18 January 2016 (UTC)

BethNaught's response to the critical question

Wikimedia should be an organisation which supports the projects it hosts. Its role is not to bring about radical changes. This is not what the many generous readers, who see banner ads asking for contributions to keep Wikipedia going, donate for. I realise this ship has sailed at the WMF but I will keep saying this.

Thanks for your comment, @BethNaught:. I believe donors also support us for our values – not just the products that the movement produces. (Independence, transparency, freedom of speech . . . ) It is not just that we have created Wikipedia, which our donors love and deeply appreciate; It is also how we have gone about it. --Lgruwell-WMF (talk) 01:13, 20 January 2016 (UTC)

@Lgruwell-WMF: And transparency involves accepting $250,000 without releasing any of the associated documentation? I smell hypocrisy. I'm sure some donors support for other reasons. But, as I was referring to, a very significant part of the WMF's income comes from banner drives and people who donate to "#keepitfree" (see Twitter). Do they know they are funding the development of divisive software, massive expansion of the WMF's staff? Do they know that the WMF is actually rolling in money, increasing its income year-on-year? Start some honest banner ads and I'll reconsider my comment. BethNaught (talk) 19:52, 22 January 2016 (UTC)

I don't think you need to reconsider your comment and I don't see us in disagreement here. You asked the question about what do our donors know. I can probably best answer this by pointing to what we tell them. Last week, as a part of the 15th Birthday, WMF ran banners in 98 languages pointing readers to the 15th Birthday site, which includes the financial data you mentioned. We will also send this report directly to our donors. With regards to transparency and fiscal responsibility, I do think WMF has room for improvement. I encourage you to share your ideas for how we could improve on these fronts here and as apart of the annual planning process, which should allow for more input from the community this year. Thanks! --Lgruwell-WMF (talk) 20:51, 22 January 2016 (UTC)

BethNaught's top 2-3 (or share your own idea)

In line with the above, I disagree with 1, 3 and 6. I endorse 2, 4 and 5 because there is still so much knowledge hidden in GLAM that we could bring out in collaboration, and the gender gap is still an issue. Efforts should be made to improve tools which assist users to translate content between Wikimedia wikis to help users who speak small languages access content natively. This promises a much bigger audience increase than rolling out a fancy new app.

Snipre

Response by Snipre 21:31, 18 January 2016 (UTC)

Réponse de Snipre à la question critique

Les communautés doivent mieux s'organiser pour mettre à disposition les aides et autres outils créés par les contributeurs. Trop d'outils sont dispersés dans les recoins de WP et sont difficilement accessibles pour les non-initiés. De plus, il y a un problème de mise à jour des outils existants, je pense notamment aux bots de maintenance, et la disparition des dresseurs de bots entraîne souvent des pertes difficiles à compenser et l'impression qu'il faut sans cesse réinventer la roue.

Un exemple classique, c'est la difficulté d'avoir une page qui recense les statistiques d'une Wikipédia et qui soit capable de fournir l'intégralité des statistiques. On perd du temps à trouver l'information, on perd du temps à maintenir, on perd du temps à remplacer quand l'outil devient inutilisable.

La Fondation pourrait régulièrement repérer des outils plébiscités par les contributeurs et mettre en place des outils similaires maintenus par une équipe technique de la fondation et qui serait vérifiés après amélioration/mise à jour du logiciel Médiawiki.

Communities must be better organized to make available the help and other tools created by contributors. Too many tools are scattered in nooks of WP and are not easily accessible to the uninitiated. In addition, there are issues with updating existing tools, for example maintenance bots, and the disappearance of bot owners often leads to losses that are difficult to offset and the impression that we must constantly reinvent the wheel.

A classic example is the difficulty to have a page that lists a Wikipedia site's statistics and that is able to provide all the statistics. We waste time looking for information, we waste time maintaining things, we waste time replacing things when the tool becomes unusable.

The Foundation could regularly identify tools favored by contributors and implement similar tools maintained by a technical team of the foundation and that would be audited after every improvement or update of the MediaWiki software.

Bonjour Snipre, Merci pour votre commentaire. Ce sont de bonnes suggestions, notre Community Tech team travaille en ce moment sur ces mêmes questions. Nous avons récemment fait une enquête avec notre communauté sure l'utilisation des outils crées par nos contributeurs. Voici les résultats de leur répo​n​ses. J'ajoute également des links pour les pages ou nous documentons ces efforts. https://meta.wikimedia.org/wiki/Community_Tech/News https://meta.wikimedia.org/wiki/2015_Community_Wishlist_Survey/Results SVentura (WMF) (talk) 21:52, 20 January 2016 (UTC)

Top 2-3 de Snipre (ou partagez vos idées)

3, 4 et 6

Mautpreller

Response by Mautpreller 21:46, 18 January 2016 (UTC)

Mautpreller's response to the critical question

The question is heavily biased so I cannot answer it. I do not agree that "knowledge needs" of users are changing in the way that is intimated here. Moreover, I do not agree that it is even possible to determine something like a "knowledge need". If you don't know about an issue, in many cases you also don't know what you "need to know". (And, from personal experience: If you learn something about an issue you begin to discover what you "need to know" in the first place; your scope of not-knowing is expanding by learning since you learn to ask new questions you never imagined before.) Even more I reject the idea that the WMF should "adapt" to aforesaid "knowledge needs". I am sure that many persons desire "short snippets" but I am also sure that Wikipedia is not suited to this desire and that these persons already have vast possibilities to find such "short snippets", especially via Google. From comprehensive experience as a longtime Wikipedia editor I say that one of the biggest problems in creating, communicating, and sharing knowledge is just this "snippets" approach.

As an education project, we should go exactly in the opposite direction. WMF should encourage and support creation and collection of in-depth knowledge, interdisciplinary knowledge, and "learning by doing", i.e. by writing and contributing. As acquirement of knowledge is invariably connected with effort, this effort should be supported instead of minimized.

"acquirement of knowledge is invariably connected with effort" -I like that a lot. I think the question here is are we going to optimize to be a quick answer engine or are we trying to be the place where people go for deeper learning. The easy answer is both, but part of the reason we do strategy is to decide what is most important, so we can prioritize resources. --Lgruwell-WMF (talk) 01:25, 20 January 2016 (UTC)

You are bringing up an important point, Mautpreller. There are two points that we need to pay attention to in this discussion: 1) The distinction between information and knowledge, 2) The way knowledge is acquired by different users. I agree with you that the knowledge needs of our users have probably not changed a lot, (the information needs may have). However, the way people learn and acquire knowledge may have also changed over the past 10 years. I think as a movement we need to be clear if we are a place to gain knowledge or information. From everything I see in the different discussions, there seems to be a general agreement that we are a place for knowledge. Going to point (2), I think it's important to acknowledge that different people acquire knowledge in different ways. Some people may start with a quick look up, then they get more curious and do an overview about a related topic, and then they may go deeper. We need to learn more about how users consume the content, for example on Wikipedia. Mautpreller, we have started this long term learning here: https://meta.wikimedia.org/wiki/Research:Characterizing_Wikipedia_Reader_Behaviour --LZia (WMF) (talk) 21:45, 20 January 2016 (UTC)

I think this is interesting as well—knowledge connected with effort. We only need to be careful that the effort is in the knowledge and learning, and not in the tools. I wonder too, if gaining trust through smaller snippets could lead to deeper participation. Heather Walls (WMF) (talk) 21:49, 20 January 2016 (UTC)

Mautpreller's top 2-3 (or share your own idea)

First: some of these approaches are dangerous and will result in corrupting knowledge instead of improving it. This is especially true for approach 6 which is downright harmful to human knowledge production and sharing. In part this is also valid for approach 1 as it focuses on the wrong thing.

My own idea: WMF should support interdisciplinarity workshops where people can try to find ways how a subject can be understood and represented from different perspectives. Interdisciplinarity is a huge problem also in science (humanities as well as natural science) because different disciplines don't understand each others, let alone practitioners and scholars. Wikipedia has unique possibilities to deal with this problem.

I do know that this is not an answer to the question asked. Rather, it is an answer to the question that should be asked: how do we improve the acquisition, communication, and sharing of knowledge?--Mautpreller (talk) 21:46, 18 January 2016 (UTC)

Edoderoo

Sources

Not sure if this is the best place, but I'll drop it here. A lot of knowledge is supported by sources/references, often online. A disadvantage is that after a few years many of those sources/online refs get moved, removed or altered. There is the Wayback Internet Archive to be of help, but it doesn't pick up all our references, and it's an external service, I'd prefer to rely on our own tools. Wouldn't it be a good idea to develop our own Reference Archive, that will store a copy of an online reference the first time it is used. Edoderoo (talk) 22:12, 18 January 2016 (UTC)

• The Community Tech team has a Community Wishlist task associated to Migrate dead links to the Wayback Machine. That may be a good place to learn more on this topic. WMoran (WMF) (talk) 21:40, 20 January 2016 (UTC)

Julius Tominius

Response by Julius Tominius 02:00, 19 January 2016 (UTC)

Julius Tominius's response to the critical question

Approach six.

Jayen466

Response by Jayen466 03:03, 19 January 2016 (UTC)

Jayen466's response to the critical question

I agree with User:Mautpreller – the way the question is framed is hopelessly biased. As it stands, it presents a strategic development of Wikimedia towards the provision of snippets as a foregone conclusion. Instead, there should be an open-ended discussion, with arguments explaining why or why not such a development might be necessary or desirable. Such a discussion could be very interesting and stimulating. As for facilitating content quality, see 7. below.

Jayen466's top 2-3 (or share your own idea)

7: Support more projects along the lines of Wiki Project Med and the efforts of James Heilman to bring content up to peer-reviewed standard. This approach should be extended to other topic areas besides medicine.

8: Collect data on edit history parameters that correlate with the quality of the information provided (e.g. number of contributors to an article, article stability, reversions, single-purpose accounts, ORES data, referencing) and make these visible to the reader in a way that is easy for them to take in, so they can have a rough guess at the reliability of the information they are consuming. Reporting such statistics also provides a visible indication of pages that might benefit from content checks and improvements.

Note: Any use of machine translation should be tightly controlled. No one should ever use machine translation to create articles in a project whose language they don't speak. Where this has been done in the past, it just caused damage. (People have been working on machine translation for half a century, and the output is unfortunately still nowhere near good enough to serve as a reference work.)

Comment: As for machine-generated articles etc., Jimmy Wales has said tonight:

'First the idea that Wikidata could be used to "construct articles" with "no need for editors to edit actual article content" is pretty absurd from a technological point of view. Major breakthroughs in AI would be necessary.'

But mw:API:Presenting_Wikidata_knowledge#See_also specifically points out:

• Reasonator and Autodesc are tools that create machine-generated articles and short descriptions about Wikidata items.

Here's an example of such an article: [1]. Here's another: [2]

These ideas may be worthwhile, but should be honestly discussed. Andreas JN466 23:57, 15 February 2016 (UTC)

MisterSanderson

Response by MisterSanderson 03:53, 19 January 2016 (UTC)

MisterSanderson's response to the critical question

...write here…

MisterSanderson's top 2-3 (or share your own idea)

Approach three: Increase content quality and timeliness by technologically enhancing our editors’ ability to create, monitor, and process content.

Approach six: Explore ways to scale machine-generated, machine-verified and machine-assisted content.

Approach five: Increase coverage in key languages through translation tools and human process.

MER-C

Response by MER-C 05:55, 19 January 2016 (UTC)

MER-C's response to the critical question

I agree this is a loaded question based on false premises ("knowledge engine" anyone?). That said, I'll ignore this and answer it anyway: The Wikimedia community curates the content and adapts it to our readers' needs. We will ask for support from you, the WMF, when we feel we don't have the tools, expertise, money or other resources necessary to do so (e.g. OSM in Wikipedia and Wikivoyage, grants, etc.). Your role is to support by fulfilling these requests and to do the things we volunteers can't, and nothing more.

Hi @MER-C: I hope I can explain why we ask this. Other editors actually highlighted this as well. We have a lot of content on our sites. The metaphor I think about is that Wikimedia is like an iceberg. A lot of the content is under water, so to say. It does not come up in searches on or on Wikipedia -- our main window to our readers. This content is good, quality content our editors create. Should we find ways to better expose it to readers? Or should we let it be?

On a related note we get a lot of requests, sometimes they are opposites. So at the WMF we have to have a level of discretion to make choices. We also don't have infinite funds, so we cannot fulfill them all. What is the best way to make those choices in your opinion? LilaTretikov (talk) 21:44, 20 January 2016 (UTC)

@LilaTretikov (WMF): Please pardon our collective cynicism due to recent board-level issues including the Doc James affair, the restricted "knowledge engine" grant, possible Google influence and the hiring of Arnnon Geshuri. Much of this cynicism comes into this discussion from the way the question was phrased -- something along the lines of "how can the WMF help editors improve our content and present it more effectively?" would be better. I agree there is a lot of obscure content on WMF sites, and because it is obscure, it is presented poorly to the reader.

To answer my own question:

• Run a consultation on novel media formats (e.g. 3D models on Commons, OSM integration, interactive stuff) that are useful for our projects. Ask us what hinders us from using existing functionality (e.g. tables, timelines, graphs, sheet music, video, sound) and how can they be improved. Back both with significant developer resources, and make them easy for us to use. (Use the Community Tech consultation as a model).
• Provide us with appropriate tools that help us identify quickly and remediate or remove users from Wikimedia communities that degrade our content (e.g. paid advocates), and help us keep them out.
• Make sure the search engine is working optimally. This includes suggesting sister project and other language content when appropriate.
• For the larger projects, shift emphasis from article creation to article improvement.
• Items 2-4 on the list.

The Community Tech survey was a great initiative and should be applied to other instances where you have to allocate development resources. For example, you can run a survey periodically on what the problems with our mobile interface are (you're not at the stage where you can ask "what new features should we add to the mobile interface?" yet), and prioritize the suggestions based on the support that they get. That you're here responding to me is also a big plus. But there are still questionable top-down decisions being made and significant resources allocated to various projects without any community input whatsoever. MER-C (talk) 08:03, 21 January 2016 (UTC)

MER-C's top 2-3 (or share your own idea)

Support 2, this is obvious.

Support 3, but only for things that we explicitly ask for. Use the Community Tech process for this.

Also support efforts to provide peer reviewed versions of Wikimedia content.

I agree that you need to support the sister projects and other languages more... once you've made significant progress against the annoying problems that affect all projects. Wikimedia is not just the English Wikipedia.

Yair rand

Response by Yair rand 07:41, 19 January 2016 (UTC)

Yair rand's response to the critical question

Support the sister projects. I don't think people are about to stop reading Wikipedia, but if they do, we already have Wikimedia projects set up for different styles of educational content. Wikiversity was built for developing (non-encyclopedic) learning resources, but it was never given the necessary technical support from the Foundation. Give Wikinews the help it desperately needs to be an functioning usable news resource. The Wiktionary communities need a dedicated team of developers to face the myriad of technical challenges ahead. Wikivoyage, Wikiquote, Wikibooks, Wikisource... These are all great projects, and they can have great futures if the WMF starts actually paying attention to their needs. Don't preemptively try to turn Wikipedia into other projects, support the non-Wikipedia communities as necessary, and the readers and the resulting communities will go where they may. And we'll be ready to receive them.

Yair rand's top 2-3 (or share your own idea)

Approach three, but not just on Wikipedia. I'm hesitant to suggest approach six, because I suspect certain people in the WMF will interpret it to mean trying to do Google-like stuff. Offer optional features to editors and communities, but don't try to push the machine-generated unreliable results directly to readers. And of course, approach seven, as explained above. --Yair rand (talk) 07:41, 19 January 2016 (UTC)

Caoimhin

Response by Caoimhin 12:21, 19 January 2016 (UTC)

Caoimhin's response to the critical question

...write here…

Caoimhin's top 2-3 (or share your own idea)

Approach six looks of most interest to me. There must be a lot of ways to automatically integrate with Wikidata, using the same data to give content in many languages.

@Caoimhin: Can you elaborate on why it is of most interest? Scale, impact, just sounds cool, some other reason? If we understand "why" better that can help us prioritize. Thanks for participating! —LuisV (WMF) (talk) 01:46, 20 January 2016 (UTC)

Lajsikonik

Response by Lajsikonik 15:03, 19 January 2016 (UTC)

Lajsikonik's response to the critical question

Focus on Wikidata. There are a lot of time wasted because of volunteers doing same work again and again, for each language separatedly. What for? Why we don't use Wikidata in similar way as Commons? Coordinates, number of inhabitants, names of mayors for towns and cities, music bands line-ups and many other "fields in infoboxes" could be easily readed form Wikidata, or even should be possible to create stubs - for some topics - automatically, for many languages, after creating a pattern from words we need for topic description. Lajsikonik (talk) 15:04, 19 January 2016 (UTC)

Lajsikonik's top 2-3 (or share your own idea)

Point after point:

1. No oppose, but I don't believe it may give a real kick
2. We should countinue to do that. Because of new content and baceuse of new users - I know that it hapenned.
3. Not most matter
4. Not possible. Who will create criteria?
5. No, no, no! Automatic translations are still far away from being quite understable, especially for Slavic languages (and also for Chinesse).
6. Where it is possible, yes. Using Wikidata.

"Suggest an approach" - perhaps we should consider grant-based system, for necessery, but repeteable and boring work. It is nothing new, and it was used, for example, for WikiMedia communicates translations. Why do not use this way to update Mairies names in > 40 k of French communes (just an example)? Lajsikonik (talk) 15:04, 19 January 2016 (UTC)

Hi @Lajsikonik:, for machine learning we are not thinking translations, but more of ORES type work. LilaTretikov (talk) 21:49, 20 January 2016 (UTC)

Ellywa

Response by Ellywa 17:19, 19 January 2016 (UTC)

Ellywa's response to the critical question

...write here…

Ellywa's top 2-3 (or share your own idea)

• I think in our society most interest is about current events, and the (scientific, historic etc) background of these events. You could enable editors to monitor real-time for instance the top-ten or top-hundred visits to articles, so the community can prioritize to keep these up to date and accurate.
• My priorities of the list would be 3 and 6 (wikidata).

Molarus

Response by Molarus 19:04, 19 January 2016 (UTC)

Molarus's top 2-3 (or share your own idea)

1. Approach one:
2. Approach six: Maybe the foundation shouldn´t explore that, but you could ask the community to do that.
3. Approach seven a: I see this as the next big thing for us: Automated Learning - What happens when computers, not teachers, pick what students learn? I think we are in the education business and Wikipedia could be used for self teaching this way.
4. Approach seven b: For short text, we could add in Wikidata a twitter-like text field with a limit space. This text could be used for a mobile and external offer. --Molarus (talk) 02:26, 9 February 2016 (UTC)

Sänger

Response by Sänger 20:14, 19 January 2016 (UTC)

Antwort von Sänger auf die Hauptfrage

Ansatz zwei: Die Zusammenarbeit zwischen der Community und Profis ihrer jeweiligen Fachgebiete (bzw. möglichst gar Eingemeindung selbiger in diese) sollte verbessert werden, inklusive Ausbau von GLAM.

Approach two: The collaboration between the community an professionals in there respective subjects (or better even incorporation of them in it) should be improved, including further expansion of GLAM.

Top 2-3 (oder teile uns deine eigene Idee mit) von Sänger

Ansatz fünf: Nicht boterstellte Stubs, sondern Unterstützung echter Autoren durch Übersetzungshilfen aus anderen Wikiprojekten.

Approach five: Not bot-generated stubs, but assistance of real authors with translation tools from other wikiprojects.

Ansatz drei: Unterstützung echter Autoren sollte immer Vorrang vor blöden Bots haben.

Approach three: Support of real authors should always trump stupid bots.

Hallo Sänger. Ich kann ein bisschen Deutsch sprechen aber nicht genug dass ich im Deutsch diskutieren kann. You are correct that real [good faith] editors are better than stupid bots. With that same token, good bots are better than stupid bots. :) Note that a project such as Wikidata can benefit from good bot activity. There are many tasks, as you know, that are tedious for humans, and easy enough for bots. If we use bots in a way that we take away the tedious job for humans, and help humans nurture what they are really good at (i.e., creativity, curation, etc.), this is a very big win for us as humans. We will learn and improve ourselves as we share our knowledge with others. --LZia (WMF) (talk) 21:58, 20 January 2016 (UTC)

Bot-driven projects, I'd call them ersatz-projects, are perhaps better than nothing, but real editors are far better. There are some menial tasks, that could be dealt with with bots, but I don't regard those WPs, who have lots of bot-generated stubs healthy and worthy projects. And I don't give a flying f*** about the health of privacy-raping, data-mining, tax-avoiding, spamming scum like noodle. They are the very opposite of the wikiverse. Grüße vom Sänger ♫^(Reden) 22:04, 23 January 2016 (UTC)

Nihiltres

Response by Nihiltres 21:19, 19 January 2016 (UTC)

Nihiltres's response to the critical question

Automated approaches to knowledge are not sufficiently robust, and not only are Wikimedia's strengths are in user-generated content, its scope is there. We should avoid approaches that require us to automatically parse meaning from content, broadly speaking (search/discovery can do some fancy things with keywords/grammar, but it should be search and not, say, "automatic Q&A system"). Wikidata's a bit of an exception, because the community's already parsed the information into structured form, but…

Nihiltres's top 2–3 (or share your own idea)

7, with significant elements from 1, 6, and 3:

Our strength is user-generated content, so we should focus on things that make it easier for users to contribute, one way or another. In particular, I think we lack pathways for shallow, "casual" contributions, and that coincides with mobile's difficulties. If we can get machines to do 90% of the work of useful maintenance edits, bringing in humans for the critical 10%, reCAPTCHA-style would be helpful. For example, there are thousands of pages in Category:CS1 errors. Many of these problems are trivial, but most aren't quite bot-solvable. If we had a mechanism for users (particularly on mobile) to, say, separate lists of coauthors into individual author-names for resolution of items in Category:Pages containing cite templates with deprecated parameters, then a bot could do the tech-oriented part of replacing |coauthors= into a bunch of |authorn= params or whatever. It'd take upkeep: continually finding and building tasks for users, plus some work to make the task robust against bad actors (e.g. reCAPTCHA's practice of confirming results against multiple users)—but it'd be worth it. It's a task that's uniquely suited to the WMF because it has the resources to code the tool and the ability to make it visible to mobile users (especially via the mobile app). The Wikidata game is a good example of how it's easily applicable on Wikidata, but it could totally be applied to maintenance tasks on Wikipedia, too.

NaBUru38

Response by NaBUru38 21:50, 19 January 2016 (UTC)

NaBUru38's response to the critical question

"Short snippets" already exist, they are called article summaries. Many of them are written poorly, so let's fix them!

Now, I disagree with automatic generation of content. We must be careful with what we publish, which means we must make that decision, not computers. Computers can't tell if information is wrong, especially because they get data entered by other people. Every edit must be verified thoroughly by people.

NaBUru38's top 2-3 (or share your own idea)

About Approach 1, uploading photos and videos must be easy. Mobile editing is useful for short facts, but it's impossible to write at length on a tiny screen.

Approach 2 refers to cultural institutions (I dislike acronyms). They are important but not because they allow to contribute faster, but because they mae contribute that others can't do. Each institution has works and experts that can't be found elsewhere.

To reduce systemic bias, we need a much more diverse community of editors. Not all of them are computer experts, not all of them have time to edit, not all of them think they are useful. We must overcome those obstacles.

Bluerasberry

Response by Bluerasberry 13:32, 20 January 2016 (UTC)

Bluerasberry's response to the critical question

The best way for the Wikimedia Foundation to adapt to changing knowledge needs of readers is to provide tools which generate good metrics of the sort that outreach organizations desire. "Outreach organizations" can mean institutions like GLAMs or it can mean community groups with any interest, but in both cases, they are sources of subject matter expertise with an interest in information distribution. Wikipedia for the foreseeable future will still interact mostly with individuals, but just the same, individuals are greatly encouraged when outreach organizations confirm that developing Wikipedia is a good use of time.

Right now, every outreach organization in the world believes as fundamental truth that they must hire Facebook and Twitter staff and that the public ought to engage with them on those platforms. In fact, Facebook and Twitter are not universal communication media, and in many ways and for many applications and fields, Wikipedia development would be a better use of their time for the sake of the audience they are trying to reach.

Bluerasberry's top 2-3 (or share your own idea)

I do not think any of these is a particular priority.

The odd part about all of these proposals, and about Wikimedia Foundation strategy generally, is the assumption that institutions will never learn to use Wikipedia and that always all organizational partnerships will originate in the Wikimedia community and operate under community oversight. It is taken for granted that institutions are so contrary to Wikimedia community values that they will not adapt them. Compare this outreach position to those of Facebook and Twitter, which both demand staff time from absolutely every organization globally. Most organizations using Facebook and Twitter do not get good value for their investment, except in the sense that they get self-satisfaction if not audience satisfaction. I think it should be a priority that the Wikimedia Foundation make it easier for organizations to get self-satisfaction by partnering with Wikimedia projects.

Organizations want a metrics dashboard. This is all. The metrics dashboard is a base to Twitter and Facebook outreach, and even when it is not used (which it usually is not), organizations get high satisfaction and comfort from knowing that if they choose to do so, they can get a good report of their engagement and impact in those projects. There is no such option to do this in Wikimedia projects. Even expert Wikipedians - like chapters and individuals getting grants - have extreme difficulty calculating metrics. What Facebook and Twitter do automatically, Wikipedians must do by hand. Insane as it is to imagine, and bizarre and otherworldly as it is, the Wikimedia Foundation requests that volunteers literally do accounting and mathematics off-wiki, calculating sums and building record sheets entry by entry, as if this were the 1990s and no one is yet certain if the Internet will ever matter. What should happen is that anyone should be able to list an arbitrary set of Wikipedia articles, or have a list of Wikipedians, and get some automated report about readers of those articles or the participation of that arbitrary group.

The lack of automated metrics is a major barrier to institutional acceptance of Wikipedia. Facebook, for example, makes a lot of money, but the value of Facebook is actually the labor invested in it by its users. All Facebook users appreciate some connection to at least some well-polished corporate feed, and those corporations invest huge amounts of labor in Facebook because they get impact reports in return and that convinces them of value of Facebook. Similarly, GLAMs would engage more in Wikipedia with metrics, but leave GLAMs aside because they mostly do not matter and have no money. With a metrics dashboard, every other institution in all sectors would be much more likely to invest staff labor in Wikimedia projects, and they would do so following all the rules if they found value in doing so. A major barrier to bringing more expert contributions of knowledge into Wikimedia projects is that the Wikimedia Foundation judges the quality of expert contributions on the basis of the willingness of those experts to arrange someone to perform esoteric arithmetic on Wikimedia data that can only be acquired using the skill set of a 1990s American Internet publisher. It is problematic that the quality of Wikimedia contributions are judged through this standard, and that the enforcement of this standard prevents participation by organizations. The unwillingness to do these math and data collection actions is also used as a basis of discrimination and justification for reputation attacks against Wikimedia contributors in the developing world.

I agree with you Bluerasberry that there is a lot of value in what you are suggesting. Basically, for every type of knowledge the user adds to the projects (upload in Commons, adding a sentence to Wikipedia, translating a label in Wikidata), we should be able to give a feedback to the user as how much value that contribution has created, and the user should be able to monitor the value added over time. This is a very big and important project. --LZia (WMF) (talk) 22:14, 20 January 2016 (UTC)

Is there really value in uploading pictures at commons? As an longtime editor I know, that most of "my" articles will not be read, but I don´t care. I don´t think that GLAMs would do so too. And even people would see those pictures, what would the value be? Money? Reputation? Or what? As a longtime editor I would say, that I have learned a lot while I wrote that articles. Maybe Wikipedia rewards editors more than readers. I remember that de:WP got a large amount of pictures years ago from a GLAM institution. The community detected lots of errors in the description of those pictures and we shared that information. That was really valuable for them. --Goldzahn (talk) 19:12, 21 January 2016 (UTC)

+1 - there is the GLAMerous tool, which could be a base for semi-automatic reporting. WMF needs to onboard this functionality. Slowking4 (talk) 22:08, 24 January 2016 (UTC)

@Goldzahn: Yes, there is a lot of value for GLAM institutions, in uploading images/videos/sounds/etc to Commons. I can't find the exact example, but I recall reading that one institution greatly increased their own website's pageviews (over the short and long term), just because of the source-link on file description pages. There are many other factors too; the best lists I can find at the moment, are at commons:Commons:Guide to content partnerships#What do you get? and at outreach:GLAM/Indicators of Success. The main point of analytics here, is to prove the value of the time spent on the work (to their managers/budget-owners), so that short-term/one-off experiments can grow into part of the regular work that they do over the years (with steadily increasing quality and detail, as everyone learns more about each other), which also become success stories for both them and us (hence memetic and even more likely to spread). HTH. :-) Quiddity (WMF) (talk) 22:06, 28 January 2016 (UTC)

I like the term "content partnership", maybe this is something WMF could work on. For example, we could give GLAM-Institutions literally data (Multilingual descriptions, Category, ...) otherwise they have to pull that out of commons, somehow, and if data is really the new oil, that should be worth something. At least, I hope so. By the way, if we could get this kind of data from the GLAM-Institutions, the commons editors would save a lot of time, I guess. Maybe a standard data format for this type of data would be useful. --Goldzahn (talk) 07:05, 11 February 2016 (UTC)

FloNight

Response by FloNight 21:07, 20 January 2016 (UTC)

FloNight's response to the critical question

Work with the broad community to develop quality metrics for content, and the tools to evaluate for improvements in quality.

FloNight's top 2-3 (or share your own idea)

• Approach Two: In addition to GLAM, regularly include STEM and health care organizations/agencies as potential program partners.
• Approach Three: Our current processes are inadequate to create, monitor, and process content for a particular topic area or topic specific programmatic activity. Those that exist are not obvious to find, easy to understand, and don't address all the needs across topic area, WMF project, or language community.
• Approach four: Wikipedia can not achieve its primary mission if the content contains systemic biases. Beyond the number and length of articles, there needs to be better tools to access article content qualtiy, including systemic bias (gender, global view, etc.) The solutions will need to be both social and tech.

Yurik

Response by Yurik 21:31, 20 January 2016 (UTC)

Yurik's response to the critical question

One picture is worth a thousand words, and one interactive visualization is worth a thousand pictures. Knowledge must be more compact - if users want quicker answers, they should not be required to read lengthy essays. Which means we should target richer content and data visualizations. We should NOT target machine-generated content, because that requires significant development resources, and our strength is in the community contributions, not fancy technology that requires large budgets. Read more here.

Yurik's top 2-3 (or share your own idea)

1, 3, 2

Milimetric (WMF)

Response by Milimetric (WMF) 21:37, 20 January 2016 (UTC)

Milimetric (WMF)'s response to the critical question

Right now the connection between our content creators and our content consumers does not exist. People reading wikipedia have no idea how the information is compiled. Celebrating content creators via projects like WikiCredit is the radical leverage point in this situation. Creators will feel appreciated and the audience will have a way to contact people and talk about what they would find useful. **Very Important**: we can't fix these problems ourselves, we just have to unite our communities and make it easier for them to interact.

Milimetric (WMF)'s top 2-3 (or share your own idea)

I support my own idea, but also 3. My idea is basically also the answer to the Critical Question. And that is, implement WikiCredit, user profiles, and celebrate content creators.

Mattflaschen-WMF

Response by Mattflaschen-WMF 21:43, 20 January 2016 (UTC)

Mattflaschen-WMF's response to the critical question

Specific initiatives are important. But so is the big picture of increasing the number of contributors and editors. If we increase this big picture participation level, these new users will come up with ways to contribute that we could never have predicted ahead of time.

Mattflaschen-WMF's top 2-3 (or share your own idea)

On specific ideas, I support:

• New approach - Supporting mobile contribution in general. This has been explored somewhat (e.g. Wikidata query game), but a lot more needs to be done to support all kinds of mobile contributions (both prose and quick microcontributions like disambiguation fixes).
• Approach three (Increase content quality and timeliness by technologically enhancing our editors’ ability to create, monitor, and process content)
• Approach one (Provide easy-to-use tools and incentives to contribute multimedia content and short-form text) - Ideally these tools will work on both mobile and desktop.

SSneg

Response by SSneg 21:47, 20 January 2016 (UTC)

SSneg's response to the critical question

The requirements for how knowledge is presented are changing. Yesterday we wanted “information”, now we want “interaction” (interactive graphs and maps, explorable infographics, playable media, galleries of files to browse etc). Tomorrow we will want, I don’t know, things to touch in VR goggles? Anyhow, Wikipedia seems to lag behind when it presents essentially text with pictures.

What about interactive explorable data charts where you can zoom into periods, toggle columns etc? Or zoomable images such as a large panorama of our galaxy or a mindmap of botanical taxons?

Creating almost anything complex in Wikipedia is complex squared. Editing tables is a pain. Importing and exporting table data requires external tools and scripts. Sorting tables is primitive and works only half the time because of number format issues etc. Creating a map is painful. Translating SVGs requires extra software. Simply putting a dot on a map is difficult and demands computer literacy above average. Even simple things like image galleries require coding and don’t look neat in the end.

Also, Wikipedia is full of incomplete data because data isn’t linked. It should be possible to input data such as populations, areas, coordinates, GDP values, birth dates, etc. in Wiki, add a reference and then reuse it across Wiki sites. I can spend 20 minutes digging up the number of employees from IBM’s 100-page annual report but I don’t have time to input it across 87 Wiki pages in different languages. Incentives such as the Wikidata game should be brought from the level of “crude prototype” to the front-page feature.

By the way, did you know that googling ‘IBM’ returns its Wikipedia article before the actual IBM.com page? This brings me to another point. Why aren’t companies interested in improving their Wikipedia pages? I don’t mean cleaning out scandals (they try to do that already), I mean contributing man hours to write deeper articles, publish archive photos into Public Domain and give researchers access to rare books or magazines that aren’t available online. They spend thousands on corporate museums and millions on corporate websites but they entirely miss the one place where people go to read about them. WMF should knock on every corporation’s door and explain the value and importance of having their company fully and nicely (in terms of coverage, illustration and data, not the tone) represented on Wikipedia. IBM’s page on English wiki gets 4000 visitors a day. I bet this is more than their Facebook page does. And all you have to do is to set a trend in California to make the rest of the corporate America and then the rest of the world follow.

Same goes for non-profits, universities, libraries, municipalities, regions, government organisations and pretty much anything else. These largest organisations have huge resources to contribute and more importantly they have first-hand access to some of the most valuable information sources and it is WMF’s job to make sure they direct those towards improving the depth, quality and timeliness of Wikipedia articles.

So to sum up, make it easier to contribute for non-technical users, help us link data points across wiki sites, engage organisations into editing or supporting editing. Then the bias will straighten itself out and the knowledge amount, depth and presentation will improve through the efforts of the community.

SSneg's top 2-3 (or share your own idea)

My selection in the light of the above: Approach 3, Approach 5.

Qgil-WMF

Response by Qgil-WMF 21:55, 20 January 2016 (UTC)

Qgil-WMF's response to the critical question

If the WMF muscle supports complex efforts to bring big loads of good content & translations, then our communities will be able to focus better on the qualitative work that partner organizations and computers cannot solve easily on their own. Careful with partnerships: other Wikimedia orgs and grassroot initiatives have done a lot better than the WMF and for a longer period. We need to keep supporting them, help them to lead, cover the gaps, instead of aiming to become the new leaders.--Qgil-WMF (talk) 21:55, 20 January 2016 (UTC)

Qgil-WMF's top 2-3 (or share your own idea)

1. Expand content faster through enabling community-led content partnership programs such as GLAM (Galleries, Libraries, Archives, and Museums).
2. Explore ways to scale machine-generated, machine-verified and machine-assisted content.
3. Increase coverage in key languages through translation tools and human process.

Trevor Parscal (WMF)

Response by Trevor Parscal (WMF) 22:17, 20 January 2016 (UTC)

Trevor Parscal (WMF)'s response to the critical question

Our superpower is that we have a community of smart people who work really hard and care about creating and curating knowledge. This is very distinct from everywhere else, where the lack of such distributed human-power has made necessary more automated approaches. The majority of what's being generated by these approaches, or even being created by individuals elsewhere on the internet, however, is information at best - not knowledge. We should stick to what makes us special and amazing: qualitative creation and curation of knowledge.

+1 @Trevor Parscal (WMF): Our unique value is human curated knowledge. And human knowledge can take different shapes and forms. We have a great long tradition of long-form text (Wikipedia) which makes us the go to source for encyclopedic content. As we consider new forms of content (audio, video) we should consider supporting communities not only in getting them access to new content (ex. historical political speeches from National Parks or 19th century political video footage from private collections) but also help organizing and connecting around topics of interest and support training and development for newer projects. I see this as a critical investment in the future of our projects. Another area to consider are smaller wikipedias, which might not have a large enough community/tools/sources access to build knowledge for that language, but if given access to translation tools and existing relevant content they could more easily scale these encyclopedias with curation efforts. Same for the newer generations of contributors who might not be as proficient with long-form text creation for example, but that are skilled at collaborative video editing, or story telling animation which could complement our existing projects. In terms of strengthening and building Knowledge I believe there is lots of room to increase support and engagement of contributors and open access to new resources. SVentura (WMF) (talk) 21:13, 21 January 2016 (UTC)

Trevor Parscal (WMF)'s top 2-3 (or share your own idea)

Approach 1 is very important, but I don't think it's impact is limited to mobile and quick lookup users, as new formats can give our content greater depth and richness. Approach 3 could help with efficiency, but it's critical to ensure that we are keeping the systems open-ended enough that the community can use them how they want and need. Finally I think that the approach not listed is to build better experiences around small groups of people working in specific areas. WikiProjects are a great start, but by further facilitating communities to form around knowledge work, users will be able to find people with similar interests who can help them learn new skills and I believe that will translate to those users sticking around longer and contributing more.

Ejegg

Response by Ejegg 22:50, 20 January 2016 (UTC)

Ejegg's response to the critical question

With English Wikipedia maturing, we can make the most additional impact by getting pages translated.

Ejegg's top 2-3 (or share your own idea)

Approach five is great, especially if we can reach out to chapters and local groups to train interested volunteers in the use of the Content Translation tool.

Jane023

Response by Jane023 17:58, 21 January 2016 (UTC)

Jane023's response to the critical question

1,2, and 3

Jane023's top 2-3 (or share your own idea)

We need to think about our current reward systems (barnstars and page hits) and think of ways to reward people for enabling others to edit

Jo-Jo Eumerus

Response by Jo-Jo Eumerus 15:08, 22 January 2016 (UTC)

Jo-Jo Eumerus's response to the critical question

I'll admit that I don't know exactly about the what the current normal is for "knowledge needs of readers" - personally I like the detailed descriptions of a topic, but that is just me. I do like Bluerasberry's ideas regarding the importance of organizational partnerships, with the note that these need to be transparent - the avoid controversies like the one around the new Trustee - and not turn into spamming or more generally an excessive control on the article content - as too frequently happens when an organization has an article on them edited by proxy.

In my consideration, technical support is something that WMF can work on. While reliable (--> as in, doesn't produce ungrammatical messes) machine translation is still a far future/artificial intelligence thing (the future dream of mine is sort of that if some information is added to an article in one project of a given language, it is automatically translated and added to other projects), a project like Wikidata may profit from having its material added by a machine. Yes, volunteers would still have to control (perhaps through input functions?) that the sources the information is drawn from are reliable and the information itself suitable, but it would remove a large amount of work that needs to be done since it's data and not formatted text that needs to be generated. Circling back to translations, machine assisted translation (sort of what BethNaught supported) is something one may want to work on. In short, these things of "semi automatic" processes where a human being needs to make the judgment-based decisions (e.g is this edit vandalism?) are left to humans but the "action" (e.g the revert) implemented by a machine.

Jo-Jo Eumerus's top 2-3 (or share your own idea)

2, 5 and 6 per what I mentioned above. 3 also is good, but make sure any such changes are well socialized - things like "Superprotect" should not happen again.

Pfps

Response by Pfps 15:20, 22 January 2016 (UTC)

Pfps's response to the critical question

I think that the best way to help facilitate better content quality is to have a central store of data to feed the different Wikimedia projects, i.e., Wikidata. The best way to improve Wikidata is to set up good common representation principles so that other Wiki projects can use the information there without having to know how information is represented in each part of Wikidata.

Pfps's top 2-3 (or share your own idea)

Six Two Seven: Improve modelling principles of Wikidata

74.65.39.235

Response by 74.65.39.235 17:09, 22 January 2016 (UTC)

74.65.39.235's response to the critical question

Partner with content experts and professionals

74.65.39.235's top 2-3 (or share your own idea)

Approach 2 should guide us.

Amgine

Response by Amgine 17:45, 22 January 2016 (UTC)

Amgine's response to the critical question

The critical question is substantially different than the questions raised by the introductory paragraph. Taken alone, the critical question addresses solely Wikipedia, while the introductory paragraph addresses how knowledge is surfaced and consumed.

Amgine's top 2-3 (or share your own idea)

5 is immediately useful. Six may be, but you need to invest the money to find out if it might be and that is too risky for an NGO. 3 is wrong: you're talking about making a better pen, but that says nothing about what it will write (or erase.) e.g. our friends in Tel Aviv use the best/more-current tools in editing en.WP content on the government payroll.

Sebastian Wallroth

Response by Sebastian Wallroth 14:51, 23 January 2016 (UTC)

Antwort von Sebastian Wallroth auf die Hauptfrage

Drei. Die Kernkompetenz der Wikimedia-Projekte; dass Alleinstellungsmerkmal gegenüber ähnlichen Projekten ist die Wiki-gestützte Zusammenarbeit von freiwillige Autoren. Das sollte unterstützt, ausgebaut und weiterentwickelt werden. Die Wiki-Technologie sollte erweitert werden. Mediawiki sollte die beste Wiki-Software der Welt sein. Und sie sollte weiterentwickelt werden, um die Überlegenheit gegenüber verwandten Technologien wie Google Docs, Etherpads, Sozialen Netzwerken (Facebook) oder Frage-Antwort-Datenbanken (Yahoo Answers) auszubauen.

Three. The core competence of the Wikimedia projects; the unique feature in comparison to other, similar projects, is the wiki-supported collaboration of volunteer authors. That should be supported, strengthened and developed further. The wiki-technology should be expanded . Mediawiki should be the best wiki-software in the world. And it should be developed further, to extend the superiority in comparison to similar technologies like Google Docs, Etherpads, social setworks (Facebook) or question-answer-databasses (Yahoo Answers).

Top 2-3 (oder teile uns deine eigene Idee mit) von Sebastian Wallroth

Sechs. Mit dem Ziel, die Botgenerierung von Artikeln überflüssig zu machen. Wenn zu einem Thema Daten und kein Artikel vorhanden sind, soll ein aus den Daten generiertes Datenblatt angezeigt werden. Dies soll aber dem kuratierten Artikel untergeordnet bleiben.

Six. With the goal to make bot generated content unnecessary. If there is data but no article for a specific theme, there should be shown a data sheet generated from the data. But this should be subordinate to curated articles.

Fil211

Response by Fil211 17:42, 23 January 2016 (UTC)

Fil211 — ответ на насущные вопросы

Сегодня проект достиг такого этапа, когда достичь возможности хотя бы поверхностного знания по всем аспектам абсолютно нереально. Поэтому в определении авторитетности и значимости источников необходимо переходить от необходимости общего консенсуса к практике формирования локального консенсуса среди авторов пишущих по данной тематике. это же касается и контроля материала на значимость и допустимость применения нестандартных форматов информации, например аудиофайлов--Fil211 (talk) 17:42, 23 January 2016 (UTC)

Today, the project has reached a stage as possible to achieve at least a superficial knowledge of all aspects of absolutely unreal. Therefore, in determining the credibility and relevance of sources is necessary to move from a general consensus necessary to practice the formation of a local consensus among the authors who write on this subject. The same applies to control of material importance and the permissibility of the use of non-standard information formats, such as audio files

Fil211 — выбранные 2-3 предпочтительных подхода (или собственная идея)

Ключевыми здесь видятся третий и шестой подходы. Например одной из проблем является сортировка новых статей и правок по алфавиту и по тематике, что позволило бы распределить усилия редакторов, сосредоточив их на наиболее удобных участках работы для каждого.--Fil211 (talk) 17:42, 23 January 2016 (UTC)

The key here seen the third and sixth approaches. For example one of the problems is a sort of new articles and edits alphabetically and by category, which would distribute the efforts of editors, concentrating them in the most convenient areas of operation for each

Chris troutman

Response by Chris troutman 12:41, 24 January 2016 (UTC)

Chris troutman's response to the critical question

I hope I misunderstand your question or we have a problem. Using some type of machine intelligence to answer questions is not where Wikipedia should be going. I also don't want to hand content to a machine so that it can parse same out to the readers. I don't support the assertion that there are "changing knowledge needs." This is an encyclopedia written and evaluated by users. If that's not what the reader wants they can go elsewhere. WMF needs to double-down on supporting the users writing the encyclopedia.

Chris troutman's top 2-3 (or share your own idea)

None of the above.

Mr. Zabej

• Approach 1
• Approach 2

--Mr. Zabej (talk) 17:17, 24 January 2016 (UTC)

Marcok

Response by Marcok 20:02, 24 January 2016 (UTC)

Marcok's response to the critical question

• 2: freeing extensive contents (both PD and free license) from cultural/historical institutions (GLAM) is a key factor to improve knowldge in all Wikimedia projects.

Marcok's top 2-3 (or share your own idea)

• 7: extensive access to reference sources (online libraries and publications) must be granted to all contributors, in order to ensure reliable references and improve contents to all Wikimedia projects.
• New and easier interfaces may be developed to use those reference sources and contents with all devices.

Sujalajus

Response by Sujalajus 22:08, 24 January 2016 (UTC)

Antwort von Sujalajus auf die Hauptfrage

Approach two

Top 2-3 (oder teile uns deine eigene Idee mit) von Sujalajus

Approach four