Talk:Wikimedia Enterprise: Difference between revisions

From Meta, a Wikimedia project coordination wiki
Content deleted Content added
Line 198: Line 198:


: Well, this not what is written in the press. For example, I have see the news here [https://www.wired.com/story/wikipedia-finally-asking-big-tech-to-pay-up/]. It is said that decision was made. And I have read in Wikinews that no discussion was done in the community. So, ''at least'', you have miscommunication problem here. [[User:Alexsmail|Alexsmail]] ([[User talk:Alexsmail|talk]]) 13:52, 19 March 2021 (UTC)
: Well, this not what is written in the press. For example, I have see the news here [https://www.wired.com/story/wikipedia-finally-asking-big-tech-to-pay-up/]. It is said that decision was made. And I have read in Wikinews that no discussion was done in the community. So, ''at least'', you have miscommunication problem here. [[User:Alexsmail|Alexsmail]] ([[User talk:Alexsmail|talk]]) 13:52, 19 March 2021 (UTC)
*:"Because the Wikimedia Foundation and chapters are just parasites on Wikimedia projects and without those who create wikiprojects, it means nothing. --sasha ([[User:Krassotkin|krassotkin]]) 07:36, 18 March 2021 (UTC)"
User [[User:Krassotkin|krassotkin]] is absolutely right. Fuck Foundation! Jimbo, you are king or moron? Act now!--[[User:1Goldberg2|1Goldberg2]] ([[User talk:1Goldberg2|talk]]) 17:48, 20 March 2021 (UTC)


== Telling us sooner... ==
== Telling us sooner... ==

Revision as of 17:48, 20 March 2021

Business development plan?

We are way past the dates in the table, but no business development plan has presented for community review yet and no update on the delay has been given. Is this project on hold? Ainali talkcontributions 22:07, 23 November 2020 (UTC)[reply]

@Ainali: Hey, less on hold and more that our timetable has massively slid. Partly due to the project readjusting a little in scope and also partly due to the capacity. We are doing another round of a focus groups and we are planning on publishing a position essay in January. I'll update the timeline with the new expected dates. Seddon (talk) 22:16, 23 November 2020 (UTC)[reply]

LLC details?

Could we get some details on the Wikimedia, LLC, please? Relevant documents, details of how it's going to function? Is it currently doing anything, does it have any employees, or is it basically a placeholder at this point? Also, this isn't stated specifically anywhere, but I assume the primary reason for the LLC's existence is because of tax reasons/nonprofit restrictions? If so, could this be stated specifically somewhere? --Yair rand (talk) 20:44, 17 December 2020 (UTC)[reply]

Dear Yair rand, I apologise for not replying to this message sooner - I took over as the 'community liaison' for this project at about the same time as you posted this message and obviously the notification of it slipped between a gap in the watchlisting of the page. Nonetheless, as you can see, today we have just published extensive new documentation that we've been working on. Yes, your assumption is correct. The specific section of the new FAQ documentation that is most applicable to your question is here.
The Foundation established a single-member limited liability company, with the Foundation as the sole member. Single-member LLCs are invisible under US tax law (you'll often see the phrase "pass through"), which means the creation of this LLC will have no tax implications for the Foundation. The LLC's activities will be reported as part of the Foundation's annual tax filings, as opposed to in a separate filing, and the Foundation will pay taxes on any taxable income earned by the LLC. However, a single-member LLC is not invisible under US corporate law. The LLC helps to insulate the Foundation from any liability associated with the LLC's activities.
Also: the LLC is a legal entity but "it" does not have employees of its own - Everyone working on the Okapi project (now formally called 'Wikimedia Enterprise') is a WMF employee. LWyatt (WMF) (talk) 15:52, 16 March 2021 (UTC)[reply]
Thank you, that's very interesting. The legal and financial structure is still not very clear to me, see Talk:Wikimedia Foundation Audit Committee/2020-07-20 for what I feel is a reasonable request. (I didn't check yet all the new pages.) Nemo 18:08, 16 March 2021 (UTC)[reply]
We intend to publish the LLC's operating agreement, and the cost-sharing agreement between the LLC and the Foundation. We're working a third internal agreement to clarify inter-entity licensing; once it's executed, we will publish all three. --TSebro (WMF) (talk) 18:27, 16 March 2021 (UTC)[reply]
Thank you, I'm looking forward to it. Nemo 06:24, 17 March 2021 (UTC)[reply]
Thank you that's also very interesting to me. As I read the essay, one of the key reason is "avoid subsidizing big corp with (WMF) donor money", but if the employees are all WMF, how is that goal achieved? 2600:1700:5B70:117F:5DFD:C0:81FE:32A 00:11, 17 March 2021 (UTC)[reply]
I'm not a lawyer and not an accountant BUT my understanding is that this will be governed by the cost-sharing agreement Tony mentions above. The simplest explanation is probably along the lines that the WMF will be able to recoup costs such as these ensuring the donor isn't left out of pocket. Seddon (WMF) (talk) 00:43, 17 March 2021 (UTC)[reply]
Can you please explain why the Wikimedia LLC is registered in Wilmington in the State of Delaware and not in San Francisco in the State of California. As far as I know the taxes in Delaware are low and so I think it is an important thing to clarify. I think that this company is registered in Delaware and that the address is Orange House is not good for the reputation of the Wikimedia Foundation. The companys address CORPORATION TRUST CENTER 1209 ORANGE ST is used by a lot of companys. I think Wikimedia LLC shouldnt have its seat at such an address. In German companys who have a seat at such a building like Orange House are called Briefkastenfirmen and the most people dont like such companys. Please think about another seat for that company. From my point of view it is important that at the seat of company it must be possible to meet a staff member. --Hogü-456 (talk) 19:33, 17 March 2021 (UTC)[reply]
@Hogü-456: Thanks for your question. I can't speak to corporate structures in Germany, but establishing legal entities in Delaware is fairly common in the United States, because the body of corporate law in Delaware is well-developed and easily understood. Using the LLC to operate Wikimedia Enterprise will help insulate the Foundation from exposure, and the clarity of Delaware corporate law furthers that objective. And, for what it's worth, I'm not aware of any analogous stigma in the United States associated with Delaware corporations.
Also: there is no requirement under US law, nor any cultural expectation, for a single member LLC's office to be co-located where an employee of the single member is based.--TSebro (WMF) (talk) 00:12, 18 March 2021 (UTC)[reply]
I don't question the legal decision here, i.e. that it's a pragmatic choice to establish the LLC in Delaware and there are legit reasons for it. However, I find it hard to believe that there is no stigma about Delaware corporations in USA, especially after the Panama Papers: there are many initiatives about the issue and even a Meryl Streep movie on the topic (granted, not the most successful) where the protagonist basically blames Delaware corporate laws for the death of her husband. (I don't endorse this movie, if nothing else because you'll probably need to accept DRM to watch it legally, but a relevant snippet is around minute 76 if you want to check the tone.) Nemo 06:00, 18 March 2021 (UTC)[reply]
Hello @TSebro (WMF):, thank you for the answer. It is good to know that the law in Delaware is clear and good to understand. I think that you should check further if there are stigmas about Delaware Corporations and I think it were good if the seat of the company is somewhere else but it can be in Delaware. It shouldnt be from my point of view at a house where many corporations have their legal seat. After what you explained me I dont have a problem with Delaware as the seat of the corporation. In Germany there are spaces for the community in different cities where they can meet other people. Maybe it is possible to establish somethig like that in a city in Delaware and then using the address of that space as the seat of the company. I think it is important that you write more about the reasons to found the Corporation in Delaware at a subpage or in the FAQ and specify if there are also tax reasons for that decision, this can be helpful to get more acceptance for the seat of the company. --Hogü-456 (talk) 19:02, 18 March 2021 (UTC)[reply]
Dear Hogü-456, just to clarify - as I think this might be a point of misunderstanding: the LLC does not have its own staff, office space, servers, or any physical "existence". All the people who work on the activities of this project are WMF employees or contractors (like me) in different places around the world. I know about the "Local K" spaces you are referring to - I have visited the LOKAL K in Köln with User:Raymond etc. - but that is a designated community meeting location supported by WM-DE. If the community in Delaware, or anywhere else in the world, wants to create a local community meeting space (like the lokal k's) they should be empowered to do that. But spaces like that are a question of priorities of local communities - what they need and want - and should not tied to specific bureaucratic processes. LWyatt (WMF) (talk) 19:19, 18 March 2021 (UTC)[reply]

For-profit or Non-profit?

Hi, just out of curiosity, is Wikimedia Enterprise LLC a for-profit or a non-profit? If it's for-profit, who are the stakeholders and who get to collect the profit? If it's a separate non-profit, will they be subjected to public disclosure compliance requirement like 501(c)3 or other similar types? Thank you! 2600:1700:5B70:117F:5DFD:C0:81FE:32A 00:13, 17 March 2021 (UTC)[reply]

The limited liability company (LLC) is a standard approach when a non-profit organization operates a for-profit activity, and will help us both manage risk and promote transparency. It is a single-member limited liability company, with the Wikimedia Foundation as the sole member. Single-member LLCs are invisible under US tax law (you'll often see the phrase "pass through"), which means the creation of this LLC will have no tax implications for the Wikimedia Foundation. That said, the Foundation is still required under US law to publicly disclose the LLC’s revenues and expenses in the Wikimedia Foundation's annual tax filings (view the audited financial reports here) as opposed to in a separate filing, and the Foundation will pay taxes on any taxable income earned by the LLC. However, a single-member LLC is not invisible under US corporate law. The LLC helps to insulate the Foundation from any liability associated with the LLC's activities. The LLC operates under the auspices of the Wikimedia Foundation, its staff are Wikimedia Foundation staff, and is ultimately subject to the governance of the Wikimedia Foundation Board of Trustees. The LLC's legal registration can be found at the State of Delaware, Division of Corporations, Entity name: Wikimedia, LLC, File number: 7828447. LWyatt (WMF) (talk) 00:28, 17 March 2021 (UTC)[reply]
Does this mean that any compensation of the LLC officers, however called, is also mandated to be in the WMF's form 990? Nemo 06:24, 17 March 2021 (UTC)[reply]
FYI, I have seen this question and have asked WMF Legal to investigate this specific issue (due to timezones it might be a while before there is a reply. This is just to notify that the question is acknowledged and I will remove this notice when a proper reply is published. LWyatt (WMF) (talk) 12:18, 17 March 2021 (UTC)[reply]
To follow up on @Nemo bis: for context, the Foundation discloses compensation for Foundation officers and select other employees in our annual tax filings, as part of our obligations as a US public charity. The IRS website provides more information about this disclosure obligation.
The LLC is a disregarded entity for tax purposes. The LLC president and the LLC board of managers are all currently Foundation employees. The roles, however, are not Foundation officer or trustee positions, as defined in the Foundation's bylaws. So, for this group, the Foundation will disclose compensation in accordance with the disclosure criteria for non-officers in our annual tax filings. And, if the LLC were to ever directly hire employees at some future date, the Foundation would then apply the same compensation disclosure criteria that applies to Foundation staff.--TSebro (WMF) (talk) 23:24, 19 March 2021 (UTC)[reply]

On the large reusers

The page mentions "a few massive companies use our projects' data", "businesses that reuse our content, typically at a large scale", "Reduce the need for high-intensity site-scraping". Sort of a curiosity, but just yesterday some FTC documents revealed that a decade ago one of them «offered remedies [...] for instance a pledge to stop scraping content from rival websites». A few initiatives we've seen around Wikimedia data might have been very different, if that had come to pass. And this has nothing to do with copyright or commercial relationships. Nemo 06:24, 17 March 2021 (UTC)[reply]

I think it is safe to say that "big tech" has no specific ideological or emotional attachment to the long-term success of Wikimedia projects, and past-attempts to replace us (e.g. Knol) show this. That is one of the reasons why this 'enterprise' service is being built. Instead of simply trying to appeal to moral arguments that it is "good" and "right" for these organisations to donate to WMF to invest in the free-knowledge that sustains their business-model... this project is attempting to build a product that the want to buy - that in no way restricts the existing options. It is a far more sustainable revenue model, and relationship model, for us if these organisations are able to "speak their language" to us - the language of commercial contracts. The fact that we have never been willing or able in the past to offer them an SLA - a contractual guarantee that the data they build their products upon (and the way that many of Wikimedia's readers interact with us) will be stable - means they have never been able to rely (in the business sense) on our existence. With a legal relationship in place, we will be in a much more strong position, a more sustainable position, to talk about things like correct and consistent attribution. LWyatt (WMF) (talk) 12:26, 17 March 2021 (UTC)[reply]

Naming and communication

Before going into the details of the discussion, I want to thank you everyone involved for this greatly improved effort at communicating and discussing the initiative properly. I understand that the WMF board will review the projects for the future based on this initial pilot, and I think that in order to make this activity permanent we'll need to see a much broader consensus to support it, as opposed to the widespread confusion (at best) that we had so far.

The new name is an improvement over the old one, because it's informative and doesn't in itself imply anything negative about the main APIs and various avenues of access. It's also a term that companies are more likely to search for.

However, there are risks in using such a recognised terminology: most other sellers use it to mean that there are two grades of services, one which is actually good and restricted, and another one which is unreliable but gratis (freemium model) or even libre (open-core model). We don't want any of that, so we need extreme message discipline. The current messaging already contains what I consider to be some mistakes, I'll point out two.

  • "Commercial reusers of Wikimedia": bad terms, the Wikimedia mission doesn't care about "commercial". We're perfectly fine with people selling Kiwix flash drives without our involvement and brands (if only!). The term also invites wrong questions such as "why don't you use a NC license then". All the sentences where "commercial" was used seem to actually care about "for-profit" reuse, which is a different problem.
  • "Enterprise-grade Wikimedia APIs, service, and support" (on https://enterprise.wikimedia.com/ ): this suggests that the main APIs are not "enterprise-grade". While that might be technically true depending in the definition (especially one based on SLAs), what people are likely to hear is that the public APIs are "less good". This is not the message we need.

More generally, it seems to me that this project is now going in the direction of a very typical selling free software business model, namely support services, and should continue in that spirit. "Enterprise-grade support [for the usage of Wikimedia data]", I'm perfectly fine with that. All the use cases mentioned in the essay are things that any third party could do if they tried hard enough, but may not want to do in-house for a variety of reasons. Placing this initiative squarely in the field of support will also help recognise that there are other actors in this field, whose offering needs to be considered if we want to ensure a thriving ecosystem. For instance, some of the use cases mentioned could be served perfectly fine by a stepped-up Kiwix offering, with some help on how to consume the HTML inside ZIM files; and I'm sure there are other MediaWiki consultants who currently provide support on how to build services which consume Wikimedia APIs.

If I understood the strategy correctly, and if that can be expressed a bit more clearly in the core documents describing this initiative, I think it will be much easier and less painful to make it a success for the broader Wikimedia movement. Nemo 06:24, 17 March 2021 (UTC)[reply]

Your first sentence here is extremely gratifying to read. You and I have been here long enough to see project-announcements/community-conversations that are less than ideal, so I am pleased to read your assessment that this communication effort is being done properly.
As for the name - and the communication of the name - yeah... Naming things, especially things that have to be understood by multiple different kinds of audiences (Wikimedia volunteers, big tech companies, tax-regulation offices), is hard! Also, describing it in a way that is legally accurate, but also brief and not-boring, is difficult. As you know, in the wikiverse, we also have this difficulty of describing to the world how WMF/affiliates are charity/non-profit organisations and the website is non-commercial, but that we do non accept non-commercially licensed content, in order to be compliant with the definition of free. Easily confusing to a journalist, a politician, or the general public!
A lot of the names and ways of describing what is now the "Wikimedia Enterprise" (formerly Okapi) project that were discussed had at least one of the problems: a) it looked like it was an 'officially sanctioned paid editing' service, and/or b) it was confusingly similar to something already available in the movement and free (especially anything to do with 'data'), and/or c) sounded like it was a 'fake' Wikipedia and/or d) undermined the legal and cultural importance of the word Foundation.
It is also important to me (and the team) that it is not a name that sounds... fun. It should be prosaic and only interesting to the specific kinds of people who might want to be customers: big businesses. It would be unfortunate if the community felt, because of a name, that they were "missing out" on a new, cool, exciting thing.
With regards to your points:
  1. yes, I will go back and look for where I've used "commercial" in the documentation and replace it with "for-profit" when it is grammatically appropriate. This might require different kinds of changes in the different translated versions depending on how languages separate these concepts in their vocabulary.
Edit: done. LWyatt (WMF) (talk) 13:22, 17 March 2021 (UTC)[reply]
  1. we have tried to ensure that the description of the 'new' thing doesn't simultaneously disparage the value of the 'existing' thing. The meaning of 'enterprise grade' is a term-of-art in the sense, as you note, that it comes with the SLA. While I acknowledge the semantic risk of unintentionally insulting the existing APIs, I feel that anyone who is actually reading documentation about APIs is already technically competent enough to know what 'enterprise grade' in this context implies.
On the topic of support services - doing some kind of paid consulting with potential/new customers of the API about how they could best integrate the service into their architecture is something that is, potentially, viable in the future. On the other hand and want to make it clear that "Wikimedia Enterprise" is not interested in getting into the area of paid mediawiki hosting, and certainly not anything that could even be remotely perceived as a paid-editing service! We have talked with the MediaWikiStakeholders group (formerly sometimes called the 'enterprise mediawiki' group) too.
-- LWyatt (WMF) (talk) 13:13, 17 March 2021 (UTC)[reply]

Edit API?

Will this Enterprise API also provide an edit API so that those big enterprise customers can also contribute data/edits back to Wikipedia projects in bulk? And not necessary new data, but also corrections/improvements to existing data? Data quality in general? I assume that many of interested Enterprise API users check data they get from Wikipedia and have some information about their quality. It would be great if they could have a way to contribute that back. Mitar (talk) 09:22, 17 March 2021 (UTC)[reply]

Dear Mitar, the project is a "read" API. Nonetheless, as it states in the FAQ under the question "Will it directly affect Wikimedia content?:
Longer-term, the Wikimedia Enterprise team also hopes to explore methods by which new information (e.g. "microcontributions") can be fed back to Wikimedia projects from the general public who are using products made by the Wikimedia Enterprise customers. This is in accordance with the movement strategy recommendation Improve User Experience which speaks of using APIs for the “...the potential for data returns”. At that time, appropriate community consultation will be made to ensure that such contributions could be sought in response to actual community needs, and in a manner that is compliant with Wikimedia editorial culture, privacy policy, terms of use, etc. LWyatt (WMF) (talk) 12:03, 17 March 2021 (UTC)[reply]
Thank you for the answer. Mitar (talk) 05:40, 18 March 2021 (UTC)[reply]

Unrelated Business Income Tax and non-profit status

This income (proxied through LLC) will probably be unrelated business income from the perspective of IRS because it is not related to the mission of the non-profit? Or is the intend to show it as related? Is there a danger that this income puts non-profit status of the Wikimedia foundation itself into jeopardy? What happens if it becomes a substantial revenue source? Mitar (talk) 09:26, 17 March 2021 (UTC)[reply]

Dear Mitar, for the first half of your question I have asked someone from WMF Legal to look at this question specifically, so I can be sure to give as precise as possible an answer.
Edit: @Mitar:, here is the answer! provided by User:Tle (WMF) - the Financial Controller of the WMF. I hope this answers your question:
Wikimedia Enterprise is a novel product offering for the foundation with its own taxability and accounting considerations. As a tax exempt organization, the Wikimedia Foundation is evaluating whether Wikimedia Enterprise is a related or unrelated business activity. The IRS has set forth the following criteria to determine whether an activity is related or unrelated: Business activity that generates revenue for the purpose of making an income; Business activity that is regularly carried out (frequency and continuity); and Substantially/Not substantially related to the tax exempt purpose, even if the income is spent on mission activities. Accordingly, if the activity is deemed unrelated and the unrelated business taxable income is required to be filed on Form 990-T - Exempt Organization Business Income Tax Return. The Wikimedia Foundation may be able to exclude some of the income that it earns from the Enterprise API platform from unrelated business taxable income.
Yes, it looks like the only criteria which is under the question (others pass) is if this is income is substantially related to the tax exempt purpose. Givens foundation's mission ("The Wikimedia Foundation, Inc. is dedicated to encouraging the growth, development and distribution of free, multilingual content, and to providing the full content of these wiki-based projects to the public free of charge. The Foundation operates some of the largest collaboratively edited reference projects in the world, including Wikipedia, the fourth most visited website in the world. In collaboration with a network of chapters, the Foundation provides the essential infrastructure and an organizational framework for the support and development of multilingual wiki projects and other endeavors which serve this mission.") I am unclear how to answer that. I am glad to read that User:Tle (WMF) believes this might at least partially not count towards unrelated business income. Mitar (talk) 06:11, 18 March 2021 (UTC)[reply]
For the second half of the question: No. As you can imagine, ensuring the charity-status of the WMF is not threatened is something that has been very carefully considered! There are some relevant pieces of information to that question in the FAQ under Wikimedia_Enterprise/FAQ#Legal. As for the issue of "what happens" if it becomes a substantial revenue source - there are specific checks and transparency rules we've described in the FAQ in the "financial" section, in particular here. It will not be allowed to supercede or replace the donations model of fundraising, as described here. LWyatt (WMF) (talk) 12:14, 17 March 2021 (UTC)[reply]
@LWyatt (WMF): does that means it was be capped (in the sense of lowering prices) were it ever to approach 49% of total revenue? As a more relevant concern, have any thoughts been put into the potential (scale of) impact on donations - people thinking "ah good, Big Tech paying for using it, that sounds like a good fiscal solution, now I don't need to debate to secure Wikimedia's running" Nosebagbear (talk) 16:45, 17 March 2021 (UTC)[reply]
Thank you for your question. I share your desire for keeping small reader donations the primary way that the Wikimedia Foundation is funded. We are in constant communication with our donors and have discussed this with many of them (both donors who give through the annual appeals and major donors). The response has been overwhelmingly positive. Our donors want us to serve more people in more parts of the world and understand that that requires revenue growth. We have a small team, Donor Relations, that answers hundreds of thousands of donor questions\inquires\complaints a year. We have only received one donor complaint regarding Enterprise API. We typically get a lot more when there is news coverage of a stance the foundation is taking or something new we are doing. We are also always running campaigns and A|B testing messages, so if there was a significant negative reaction to this, we would see it relatively quickly. We care about how our donors will react to this, we are proactively talking to them, and we are carefully listening to their response. To the second part of your question, from all of the customer discovery conversations that we have had, we do not believe Wikimedia Enterprise will produce revenue on the same scale as our reader donations, which still continues to grow every year! However, if we are wrong on that, we are in control of how much we invest in growth of Enterprise API vs. growth of reader donations. There are lots of different course corrections we could make (if and when). We are excited to one day have Enterprise be a supplemental revenue stream for the movement’s growth, but we are committed to keeping the majority of financial support coming from reader donations. --Lgruwell-WMF (talk) 17:45, 17 March 2021 (UTC)[reply]
Hm, FAQ touches other potential issues. Here I am really thinking only about Unrelated Business Income. There is no official limit, but a too high unrelated income can bring issues with IRS and potentially losing the status. This site mentions a rule of thumb of 20%. Mitar (talk) 06:06, 18 March 2021 (UTC)[reply]
"We are also always running campaigns and A|B testing messages, so if there was a significant negative reaction to this, we would see it relatively quickly": this is an extremely dangerous attitude to have, see w:en:McNamara fallacy. I urge you to abandon it at once, Lgruwell-WMF. I'm available to suggest more pointers on why, if you're interested, but I'll start by mentioning Hannah Arendt. :-) Nemo 06:12, 18 March 2021 (UTC)[reply]

Non-general funding

I had originally considered this humourously, but on more consideration, I think that just adding it to general funding is not wise. I do appreciate that general funding was a logical opening choice for the team to propose.

Funding should go:

  1. To it's operating, and future development - to whatever amount is suitable
  2. Potentially to a LLC-specific reserve fund, not to rise higher than 6 months of operating costs.
    Of the remainder (net profit-reserve build)
  3. 30% is to go to efforts to technological efforts: mobile editing, Core (long underfunded), and efforts to acquire new editors (Growth Team, IP, etc)
  4. 40% is to go to general funding, as anticipated
  5. 30% is to go to Community Tech, either funding a vastly larger tech team, or a spun-off second team allowing tasking to different uses (e.g. one for Wikipedia/Data/Commons, one for other projects etc)

The WMF would need to commit to maintaining their current levels (inc. inflation) of funding for those purposes that come from normal donations, to avoid fund being shifted to general funding by the back door.

Smart assistants and search engines resolve queries for people without taking them to articles. That causes two issues. The Enterprise project looks like a good resolution for the fact that most donors donate from visiting the actual site. HOWEVER. The other issue is that most new editors come from those who visit the actual site - we need to fix that problem to avoid long-term structural goals.

The Community is also nervous about what will feel like many as a step away from our inherent nature. It's not an unreasonable project by any means, but it does come across as...uncomfortable. Beyond that, the Community has also registered concerns about how the WMF uses funds in non-ideal fashions. I think a more targeted funding model would help alleviate concerns. It brings its own issues (not least, feeling somewhat like requesting a bribe as I write up the idea) but I think it moderates the biggest issues and gives a chance for a single project to combat two category 1 issues - those which could, in the long-run, risk the viability of the project (it needs editors and it needs money). That's a goal worth reaching for Nosebagbear (talk) 17:23, 17 March 2021 (UTC)[reply]

Dear Nosebagbear - I appreciate the level of detail and thought that you've put into this reply and apologise in advance that, in an attempt to respond to you as soon as possible today I might not be as detailed in reply.
For the way that the revenue generated should be 'split' into different groups/proportions/priorities - there's two things we can state unequivocally: that the service will be self-funding, and that anything beyond that will be used for/by the movement. Options include all of the things you've mentioned and, importantly, the Endowment fund. Beyond that, however, I would not personally be willing or able to comment - because those are strategic decisions about movement funds and should take place at a "higher" level than within a specific project. By comparison/equivalent: The team that organises the fundraising via emails does not, and should not, get to decide how the money they raise is spent. That would be inappropriate. Equally, the team involved this project should not have any specific control over how the money is spent - it is movement funds and should be allocated according to movement priorities. [I was a member of the Resource Allocation working group (as a volunteer), so I care about that issue!]. Legally, where the revenue from this service is allocated is ultimately the responsibility of the board of trustees. More broadly, the whole movement strategy process (including concepts like the Global Council) will perhaps create new modes for deciding on these kinds of resource-allocation questions, but that is beyond the scope of this project and this talkpage.
With regards to your point about voice-assistants, the fact that a not-insignificant portion of Wikimedia's "readers" (listeners?) are not visiting the website to access the information is both good and bad - as you identified. It means we are getting the knowledge to where people need it, but also that fewer see the fundraising banners, and equally, that fewer are able to potentially contribute to the site. We propose a longer-term goal of finding ways to bring people "into" the community in some manner in this section of the FAQ. Diversifying the revenue to ensure sustainability in the face of the potential slow loss of "eyes" on the website is precisely what this project is directly intended to achieve (see "Why Charge?" in the essay).
As to the inherent feeling of 'discomfort' - I completely understand, which is partially why I and others in the team have spent so much time to craft, test, improve, craft, test, improve - all this documentation and the various promises and principles within it! -- LWyatt (WMF) (talk) 20:25, 17 March 2021 (UTC)[reply]
@LWyatt (WMF): it's certainly beyond the remit of this team to decide the use of funding it generates. However, that is a question that is relevant to whether this project is viewed as a problematic but ultimately beneficial change, or a harbinger of a change in the WMF's focuses. So it's not just a case of "handle project, then handle funding in next funding review", as we can't step back should the latter come round with a non-desired outcome. It would, I assume, require even your project board (not that I'm sure of the WMF's current project management methodology/structure) bump it further up the chain.
With regard to the principles, is there a full list of considered occasions where "full honesty and transparency" would not be possible. I can think of a couple, but this is an instance where we don't need example cases, we need something closer to a list of likely exemptions before they come into being Nosebagbear (talk) 14:38, 18 March 2021 (UTC)[reply]
The question of "how transparent" (and the inverse: "how much privacy") goes as far as what is in our own control: for example the code is open source (tech transparency), the policies about personally identifiable information are part of WMFs existing policies (user privacy). The finances are at least as transparent as the WMF's existing processes because all of this project must be reported as part of the WMF's "form 990" and audited financial documents. What we can't promise is transparency of behalf of others. We only control what we control and it is beyond the scope of this project to force other organisations to publicise their technical, legal, or financial documentation.
As for the oversight structure (especially since the contracts with 'customers' are owned by an LLC) there is a board, all of whom are senior employees of the WMF in their capacity of those executive functions, and therefore responsible to the mission and overseen by the WMF Board of Trustees. LWyatt (WMF) (talk) 15:22, 18 March 2021 (UTC)[reply]

Relation to IP masking

Hi, at the same time this is happening, another WMF team proposes to mask IPs: IP Editing: Privacy Enhancement and Abuse Mitigation. As far as I understand it, regular users and readers will not be able to see it but some users with additional rights (functionaries like checkusers, stewards, etc.) might be able to see it. Is this covered by the principle No non-public personally identifiable data. The information provided through the Enterprise API service is the same which is available publicly on Wikimedia projects.? Will Enterprise give access only to masked IP addresses (if this will be enabled)? —DerHexer (Talk) 17:30, 17 March 2021 (UTC)[reply]

I would imagine that as with API draw at the moment, all information is drawn as if an IP editor were looking at it Nosebagbear (talk) 17:35, 17 March 2021 (UTC)[reply]
Dear DerHexer, The simplest answer is to reiterate the principle "No non-public personally identifiable data. The information provided through the Enterprise API service is the same which is available publicly on Wikimedia projects". But, to be explicit about the question of IP addresses: customers of the "Enterprise API" only get the same data that is otherwise publicly available. Enterprise API customers do not get additional, privileged, better, or in any way 'different' access to any kind of data (personal, or content etc) than is available to any normal user on the website (or user of the existing APIs). Rather, the product is built for higher speed and volume of content flow, and service-support for that.
Usernames and IP address can be considered personally-identifying information, yes, but Enterprise API customers get the same personal data as current users of any of the existing APIs would get. With the IP-Masking project specifically: However the existing APIs would change as a result of that project's results - the same would also occur with the Enterprise API. Perhaps that is too general/non-technical an answer for you? If you want more specific details of a specific technical aspect, please advise what in particular, and I can find the relevant expert to give a response here. LWyatt (WMF) (talk) 18:21, 17 March 2021 (UTC)[reply]
Hey @DerHexer:. I'm the PM for the IP Masking project. I would agree with what LWyatt (WMF) said above. With IP Masking in place, IPs for unregistered editors would be considered personally identifiable information and would not be available for public eyes. They would be treated similarly to how we treat IPs for registered users - only accessible to a limited set of users. I hope that answers your question! -- NKohli (WMF) (talk) 12:27, 18 March 2021 (UTC)[reply]
Yes, thank you, that helps. Best, —DerHexer (Talk) 20:20, 18 March 2021 (UTC)[reply]

Business case examples of use

I just found out about this project after another user posted a question on the help desk about this Wired article [[1]]. It's paywalled but I'm interested in understanding what some real world examples of business use cases are. Perhaps a "Business cases" section could be added to the content page, with a mixture of real and hypothetical examples of how companies might use Wikipedia Enterprise services? Or maybe at least a few examples in the Wikimedia Enterprise/FAQ#Financial section? Cheers. Timtempleton (talk) 19:55, 17 March 2021 (UTC)[reply]

@Timtempleton: I'll raise this with the team. We definitely have lots of thoughts on how and where we can be used, based on real world examples. In the case of a digital assistant or search functionality our content is integrated within a broader ontology linking multiple data sources, frequently referred to as a “knowledge graph”. The knowledge graph can then be leveraged to provide additional context from related topics or concepts by products. Queries from users are interpreted via natural language processing to identify a particular topic or subject matter. The ontology or knowledge graph is polled for information and the response is relayed back to the user. That applies in any search based product. Simpler integrations of Wikimedia content often involve context provided programmatically in map products, used to counter disinformation on social media and integrated with streaming platforms. Seddon (WMF) (talk) 23:17, 17 March 2021 (UTC)[reply]
@Seddon (WMF): Interesting. I'm most familiar with the Google Knowledge Graph, particularly since we answer a lot of questions at the Help Desk and Teahouse from people upset that the graph is showing wrong info, such as the wrong person's photo (which is of course a Google issue, not Wikipedia). Did you see this [[2]]? The paper mentions Stanford's DeepDive project. [[3]]. The computer science is beyond my level of understanding, but I wonder if it might have applications for the Wikipedia Enterprise project? Timtempleton (talk) 23:38, 17 March 2021 (UTC)[reply]
@Timtempleton: that one is new to me personally, but not sure about the team. The ontology and graph projects I'm most familiar with are yago, dbpedia and babelnet. Seddon (WMF) (talk) 04:16, 18 March 2021 (UTC)[reply]

Firewalls and undue influence

Has any thought been given to some type of firewalls, such that those companies which are paying substantial sums for data access cannot exercise any undue influence (read: any influence) over Wikimedia project content by threatening to stop using the service? Is there a plan in place to ensure that if they try that, they really will be told "Well, that's your choice, and we'll be sorry to see you go", even if they're spending millions per year? Seraphimblade (talk) 22:57, 17 March 2021 (UTC)[reply]

@Seraphimblade: Non-profits should in general do what they can to avoid the problem of a single dominating revenue stream. From a governance and fiscal perspective, the best way to counter to that is ensuring a diverse & sustainable resource base and a healthy reserve. Diversification is something that the Wikimedia Foundation has been working to improve over much of the last decade. We used to be reliant on banner fundraising on desktop. Over time we've grown mobile fundraising, established an email program, built a major gifts program, setup an endowment and recently started expanding our planned giving/bequests program. Wikimedia Enterprise is an additional feather in the movement’s cap in that regard. Having and maintaining that diversity in terms of funding offers us protections from the volatility of any given single revenue stream, which also includes Wikimedia Enterprise.
The endowment acts as one way to reduce long term volatility in funding, turning one time or large amounts into a long term sustainable source of funds and it would not surprise me if some component of the money from Wikimedia Enterprise goes towards the endowment. It will happen implicitly anyway as part of the foundation's annual budget even if an explicit component is not tied to Wikimedia Enterprise specifically. The second means of reducing any funding volatility is through having a healthy reserve. I know it's sometimes derided by the certain members of the community, but the Wikimedia Foundation having around 18 months of reserve is what I would consider to be best practice for a non-profit of our size. I would also expect to see our reserve keep pace with any changes in our budget. It should more than permit us to absorb any posturing by a commercial entity that might try to cancel a contract.
As we have noted, having this activity undertaken by a limited liability company means the Wikimedia Foundation is offered better protections from the various liabilities, business activities entail. So even if a company were to take aggressive business practices, this would firewall the Wikimedia Foundation. The endowment moving to its own 501(c)(3) similarly firewalls the movement's future survival and so there are multiple layers of risk mitigation in play covering both the short and long term for the movement.
Fundamentally though I consider the editorial processes on the Wikimedia projects and the independence of communities in that respect to be sacrosanct. Interference by the Wikimedia Foundation whilst really not feasible is also fundamentally undesirable. It would expose the Wikimedia Foundation to huge risk and undermine the section 230 protections. (which on a side note highlights the importance of section 230 to the communities independence and not simply the WMF as a host).
So when you combine diversity, sustainability, and stability in a fiscal policy like the WMF has, it acts to reduce risk across the board and applies to the revenue generated by Wikimedia Enterprise. That existing fiscal approach combined with the need for editorial independence, not just by the community but also the Wikimedia Foundation, means that our ability to say “sorry to see you go” isn’t just going to be a bluff. Hope this helps. Seddon (WMF) (talk) 04:04, 18 March 2021 (UTC)[reply]
Thanks. Yes, it is good to know that it has been considered. Seraphimblade (talk) 06:53, 18 March 2021 (UTC)[reply]

Please Stop

I am forced to testify that most Wikimedia contributors had known about the commercial side of this project only yesterday. Some from the press release, and some from the news at all. It means only one thing. We have completely disrupted communication between the communities and the Wikimedia Foundation.

Maybe that's why I heard a lot of harsh statements from community members about this project yesterday.

Therefore, I propose to explicitly and defiantly stop and hold a broad discussion with the project users. Here on the Meta on separate page.

With noticeable notification of the users of each project and with the opportunity to speak out in their language.

And completely stop the project if no consensus with the community is found.

And I suggest that we always do this. We need to have a preliminary discussion with the communities on every major decision.

At least so that the project users haven't known about it from newspapers. Ideally, we should reach a consensus with project contributors on each principle issue.

Because the Wikimedia Foundation and chapters are just parasites on Wikimedia projects and without those who create wikiprojects, it means nothing. --sasha (krassotkin) 07:36, 18 March 2021 (UTC)[reply]

@Krassotkin: I disagree with your assertion that this is a surprise that has come out of nowhere. To quote User:Guillaume (WMF) from the wikimedia-l mailing list, who put it much better than I could:

From the Strategic Direction in 2017 ( https://meta.wikimedia.org/wiki/Strategy/Wikimedia_movement/2017 ), to the WMF revenue strategy in 2018 ( https://meta.wikimedia.org/wiki/2018_Revenue_strategy/Summary ), to the working group on Revenue Streams in 2019 ( https://meta.wikimedia.org/wiki/Strategy/Wikimedia_movement/2018-20/Recommendations/Iteration_1/Revenue_Streams/1 ), to the final recommendations in 2020 ( https://meta.wikimedia.org/wiki/Strategy/Wikimedia_movement/2018-20/Recommendations/Increase_the_Sustainability_of_Our_Movement ), and now to this, it's certainly been a long journey, but it's great to see this coming to life. Taking this idea and shepherding it towards implementation while respecting the ethos of the movement is not an easy feat, and the team really seems to take this to heart, which is a good sign.

This project has been in progress for a few years, with multiple opportunities to comment along the way. In fact, you could talk with your community and add your comments, questions, and concerns now. The staff working on this seem highly responsive to feedback and questions. Given the previous and ongoing consultation here, asking them to stop the project does not make sense to me. --Deskana (talk) 13:43, 18 March 2021 (UTC)[reply]

  • Most community members are not active on meta, and to be honest I don't remember any notification in previous years on my home wiki about this project at all. I also heard of this from the news. and kinda taken aback, this sounds like WMF essentially sells the data volunteers worked on for free. Xia (talk) 14:56, 18 March 2021 (UTC)[reply]
  • You know, a typical pattern for 99% of wiki contributors is writing and improving articles (uploading images and the like). The vast majority of Wikimedia users do not participate in discussions on the very small community that has developed around the Wikimedia Foundation and enjoying its benefits. Moreover most Wikimedia users don't speak English and cannot participate in these discussions at all. In fact there are essentially two disjoint Wikimedia movements. The first ones to create wiki projects and make money (through donations). And the second, who spend this money on some strange projects. That's why the communities don't know about Foundation's initiatives and reject them. We must put an end to this. Every Wikimedia project contributor must participate in decision making. Nothing should be accepted without prior and explicit approval from those who create the wikis. We have to start right now. --sasha (krassotkin) 15:03, 18 March 2021 (UTC)[reply]
Dear krassotkin, I appreciate that you have taken the time to come to this page on Meta to express your concern about this project. We can, of course, also continue the discussion on the Russian WP village pump or back on social media. Nonetheless, it is useful that you have shared your position in this central place as others (like user:Xia just above) may share that opinion). As I understand it, there are two, overlapping concerns here:
- About this project specifically (concern about the assumptions behind it, the design/implementation, the value to the movement, the potential for it to undermine our values...),
- Regarding the sense of being forgotten, or left-out, of decision making processes.
Is that a fair (but simplified) summary?
Just for sheer practicality - it is impossible for every person to be personally consulted about every topic. In fact, there is often the reverse criticism: that the Wikimedia Foundation is sending too many surveys and "come and discuss xyz..." messages. Just because of human nature - some people will want to be more informed, sooner, and some people will want to be left-alone, and that will change depending on the topic. For this project specifically: The public pages for this project (on Meta, MediaWiki, Phabricator) were started in mid-2020 with what was - naturally - relatively little detail since there wasn’t anything “built” yet. But nonetheless there were threads on mailing lists, phabricator tickets etc. at the time too. Now, at this stage of development, there is actually something to “show’ (both in terms of policy and technology) that is concrete enough for people to give useful, actionable, feedback and commentary upon. Much earlier and it’s all hypothetical, much later and it’s already ‘done’: and in both of those cases it would be rightfully very frustrating to be asked ‘what is your feedback?’ of an either non-existent, or an already-finished, project. We have tried to get the ‘Goldilocks’ moment in the middle of those two extremes: But it’s never going to be the right moment for everyone simultaneously.
However, here is one, practical thing I can offer you:
I promise that the documentation pages for "Wikimedia Enterprise" (the Essay, FAQ, Principles, and Main page) will all be translated and published in Russian, next week.
The information was all published on Tuesday in English, French, German, Spanish, and some Italian. I will come to the ru.wp Village Pump (and also the Wikinews article you linked to) to directly inform the community there, when it is published.
I realise that this does not solve your frustration about hearing about this project later than you want. But it is the only thing I can do now to help increase the sense that you, and Russian Wikimedia community's feedback, is being sought. LWyatt (WMF) (talk) 16:40, 18 March 2021 (UTC)[reply]
You use the perfectly right word when you write : "the information was published" : Top-down only, unfortunately. --Croquemort Nestor (talk) 17:11, 18 March 2021 (UTC) PS : sorry I don't go so often to Meta or Phabricator (!)[reply]
Through the latter half of 2020 the team had 1:1 community interviews and roundtable discussions with a variety of people across the movement. The input, feedback and questions we received as part of that and from ongoing feedback were hugely valuable to the team and the project. Those conversations have already heavily influenced our work and it will continue to do so. If someone wants to speak to the team we’ll jump at the chance. We are attending the SWAN meeting this weekend, running two office hours tomorrow and will hosting the next Wikimedia Clinic. We’ve also engaged in as many community spaces as possible both on wiki and across the numerous communication platforms the community uses. We are still very much at the beginning of this project. Seddon (WMF) (talk) 17:17, 18 March 2021 (UTC)[reply]
  • Liam: I am grateful to you for your openness and desire to explain the project to the widest possible range of wiki projects contributors. But I'm writing about the need for a broad discussion not for myself. You need this discussion. Only then will you see how much the creators of wiki projects do not accept the idea of turning Wikipedia into a business project. We will not be able to attract new members. We will not be able to convince contractors to transfer their materials under a free license. And so on. Only because in the eyes of society Wikipedia will cease to be a charitable project, but will turn into a business. Whatever you write in the explanations the words business, commerce, sales, and so on are toxic to us. That is why we have to bring up only one question for discussion. Are you ready to turn Wikipedia into a commercial enterprise? And I know the answer. Just ask our communities. That's why we must come up with another scheme for the implementation of this project. --sasha (krassotkin) 17:19, 18 March 2021 (UTC)[reply]
  • @Krassotkin: I don’t see how providing an enterprise-grade API, which provides the exact same data as the existing APIs, turns Wikipedia into a commercial enterprise. Large businesses have different operational requirements from more casual API consumers; higher query throughput, higher uptime, optimised response format, customer support with guaranteed response times, to name a few of off the top of my head. Nothing about those things contradict the Wikimedia movement mission or its commitment to openness and transparency. Securing a revenue stream from this to continue funding the movement, whilst keeping the core API free, open, and accessible, seems entirely consistent with the movement’s mission to me. Deskana (talk) 10:37, 19 March 2021 (UTC)[reply]

@Seddon (WMF): what is not understood by WMF (or you, not sure) is that simply posting things on meta and then saying, well it was on meta for a long time is a very poor excuse You are well aware that most active users don't even know what Meta is. Let alone Phabricator. Especially in Wikis where a lot of users don't speak English or have poor understanding of English. Not consulting the wiki communities and simply expecting them to browse Meta for new project announcements, like, every month is really.... disappointing. This feels like WMF wants to do something, does it anyway, and if people complain they were not consulted, their complaints are dismissed by simply point them to discussions on meta. Yes, you should send invitations to such discussions to every village pump, and if people are not interested, they won't come, but at least will know there is something going on. And yes, WMF sends messages for all the unimportant minute technical changes of the software nobody reads or cares about, but it seems like that they don't send such messages about important stuff like this. Feels like you are being selective a little, no? If I may sound sarcastic that is because this is really truly disappointing. Such a large scale thing, that fundamentally changes Wikimedia as a whole, and most Wikimedians get to know it from the news. Simply because they are not frequent meta users. Thumbs up! Xia (talk) 17:37, 18 March 2021 (UTC)[reply]

@Seddon (WMF): I am not active wiki-user of couple of years, but I was one in the past. This is just outrages! Such thing should be at least discussed at Wikimania. It is not enough to publish them on Meta. And what is Phabricator? I will go to check... My personal opinion that is should be holded immediately until at least next Wikimania. In the past you have bot that goes over user pages and for every 'active' user (at least 10 login in the past month, for example) posted some message such as "It is Russian week coming! You can participate..." and link to the page. Or "Do you mind to make article patrol?" Something along these lines can be done across whole languages. It is ok to provide link to meta and has discussion their. But you must have such discussion first. Alexsmail (talk) 18:22, 18 March 2021 (UTC)[reply]

Wikimania is actually pretty poor as a discussion forum and I wish the WMF would use it less, as it biases discussion towards senior members of affiliates, those with income, and those with the inclination towards spending large periods of time at something like Wikimania. Certainly I'd feel using it as a some "any major discussion must include 1 Wikimania in its timeline" as a very poor idea. I do think that the various WMF teams need to get together and put together a 250 word monthly report, with links, that as well as being on meta, is translated broadly and dropped to at least the 15 biggest wikipedias primary community forum. That's not the Enterprise team's fault for non-implementation (it has been requested repeatedly since at least as far back as the strategy "consultations", and that's just what I got into meta - could well be before), but the Community Engagement folks could save themselves long-term aggro if they handled it. If they got a couple of volunteers who could translate from those communities it shouldn't cost anything but about 3 hours of WMF time a month and maybe another 3 in Community time. Nosebagbear (talk) 00:58, 19 March 2021 (UTC)[reply]
In the past I have been on my wiki-meating and one Wikimania event. While on wiki-meating we generally discussed some of our local issues, it was great to meet in person people that you're cooperative with and sometimes it was great place to settle down the conflicts that has been occured. It was also a good place to educate some less experience users. But it was bad place to even talk about some globall wiki-related issues. But Wikimania is other thing. The number of things that I've learned their was enormous. And I do remember announcement of couple of initiative to be discussed that was done their. After Wikimania ended I do remember the spark of discussion on many Wikipedia's. Wikimania is indeed bad place to make a decision, because of the biased population, but I think this is good place to make announcement of such big initiatives. I want to clarify one more point: may be eventually this is good decision, but it can not be done without proper discussion. Alexsmail (talk) 13:45, 19 March 2021 (UTC)[reply]

Hello, I would like to thank everyone for taking the time to read the update and share your thoughts. That is just what the announcement this week was meant to do–provide details to enable informed discussion about this topic, while the project is still young. I would like to add that I agree posting information on Meta alone will not reach all community members or bring a deep enough discussion. I also don't think discussing new initiatives at in-person events is enough either.

I can share a little extra background context, which led to the announcements you read this week: I was a member of the Revenue Streams movement strategy working group. In 2019, there were a few in-person meetings where creating a new revenue stream for the Movement with a new API, without changing the existing APIs, was discussed. At the March Wikimedia Summit 2019, “monetizing an API” was discussed within the working group, as well as with members from other groups. At Wikimania 2019 in Stockholm, the revenue working group presented an update on the recommendations, including the so-called “paid API”. I participated in a large group discussion as well as many smaller discussions throughout the event. In the second half of 2020, once the team for this project was assembled, they had conversations with ~70 Wikimedians from across the moment to design the operating principles that are now available for comment. There are several “office hours” this week (listed at the top of the Wikimedia Enterprise page), and the team welcomes any specific requests to meet. If there is a critical space where you think we are missing out, please let us know or invite us to participate. MeganHernandez (WMF) (talk) 13:35, 19 March 2021 (UTC)[reply]

Well, this not what is written in the press. For example, I have see the news here [4]. It is said that decision was made. And I have read in Wikinews that no discussion was done in the community. So, at least, you have miscommunication problem here. Alexsmail (talk) 13:52, 19 March 2021 (UTC)[reply]

User krassotkin is absolutely right. Fuck Foundation! Jimbo, you are king or moron? Act now!--1Goldberg2 (talk) 17:48, 20 March 2021 (UTC)[reply]

Telling us sooner...

Look, I really like the WMF. There are a lot of people I respect and admire that work for them, and they do a lot of work that goes unnoticed most of the time. I also like this idea even if I do have some concerns about the OpenStreetMap case study.. or if this API could indirectly contribute to machines of war and state.

I just want to know why we couldn't have talked to sooner about this. English Wikipedia made you all a forum for you to discuss things with us, yet it seems community members use it more than community liaisons. I'm not trying to blame LWyatt (WMF) for not posting there, but shouldn't that be a part of the job training for a community liaison?

I must be missing something here for why I always feel so out-of-the-loop. I know I am some nameless loser who edits Wikipedia as a hobby and all, but I still thought I was part of a global community where I have an actual voice in how things operate. Now I just don't know if it even matters what I say.. :( –MJLTalk 21:21, 19 March 2021 (UTC)[reply]

LWyatt (WMF) gave a really detailed in the office hours. I'll link to it when it gets posted on Wikimedia Commons. –MJLTalk 22:26, 19 March 2021 (UTC)[reply]

Letting the wolves in at the door

It's one of the oldest dictums out there - money is power. If you control the purse strings, you can control everything else. Or, to put it another way, a budget document is a values document. I believe that the WMF is creating this in good faith. After all, why pass up the potential for a lot of money from some of the world's richest companies if all it's doing is affecting the data flowing out of Wikipedia, and won't impact what content is added or deleted?

That might be true today. But the WMF is putting itself in a dangerous position. If the revenue stream is as successful as the WMF hopes, at some point, it will make up a substantial portion of the WMF's budget. And then the big companies will be in a position to lean on the WMF for changes, in big ways and small, obvious and subtle. Wikipedia is an idealistic place. It will be destroyed by the rapacious maw of capital if it is not carefully managed. The WMF and the community have kept the wolves at bay for years. But this lets them in at last. One step at a time, $$$ will change what Wikipedia is. This idea is a mistake, and a dangerous one to this project. Ganesha811 (talk) 21:26, 19 March 2021 (UTC)[reply]

@Ganesha811: you might find my above answer to Seraphimblade addresses much of your point. Seddon (WMF) (talk) 21:38, 19 March 2021 (UTC)[reply]
Thanks for pointing me to this responnse, Seddon (WMF). I think it's a reasonable argument - for now. Assuming good faith is key here, and I really do believe that the current WMF is approaching this cautiously. But will it always? It's a very rare organization indeed whose culture improves over time. The WMF is made up of people - but what happens if the type of people at WMF change? A good organization is not only committed in the present to doing the right thing, but is actively future-proofing itself against internal mission rot. Wikipedia has never sold anything before (other than some t-shirts). In doing so, it crosses a fundamental line. Slippery slope arguments are often overused, but this is one case where they apply. Future versions of the WMF may not be as careful on these dangerous, slippery cliffs as the current one appears to be. Ganesha811 (talk) 21:49, 19 March 2021 (UTC)[reply]
@Ganesha811: just an extra note. One of things we mention in the FAQ is that actually we have done something very much like this before. Paid data services were offered since shortly after the WMF's inception, providing feeds to enable third-parties to host their own local databases. The creation of this service was what led to the initial hire of Brion Vibber who was the very first employee, and was used to help bootstrap the Wikimedia Foundation in the early years. The service was closed to new customers in 2010 and the service was finally decommissioned in 2014 mainly due to lack of maintenance. Seddon (WMF) (talk) 22:07, 19 March 2021 (UTC)[reply]
  • This is a funding issue, so the project team can't really answer it which is why I keep asking them to grab someone from further up the tree, but the only way to ultimately prevent a slippery slope case is to cap the max influence they can ever have on the project. The principles, by using the word "supercede", should cap the funding % that can come via this channel at 49% of total revenue. However that is way too high. I would say that the WMF should commit to capping the total revenue that can come in through any paid channel at, say, 20-25%. When it runs over, then prices are lowered. While this would obviously reduce the max scale of available funding, it prevents any pressure ever being exerted. I would not expect to see that pressure come in the form of demands for content changes and so on. I would expect to see it in the form of pressure for development or focus in areas outside the purest definition of this LLC's scope. Nosebagbear (talk) 01:19, 20 March 2021 (UTC)[reply]
Just as a point of history, it is not true that WMF has never sold anything before. A very long time ago (I'm old!) we did a deal with Answers.com to provide them with an API/feed for a fixed monthly price. It wasn't a super successful project and nothing particularly good or bad came of it.--Jimbo Wales (talk) 10:18, 20 March 2021 (UTC)[reply]
Fair enough, Jimbo Wales, that's a piece of history I didn't know about. I'd be curious to hear your thoughts on future-proofing the WMF against internal cultural change to make sure that this revenue stream never results in problems. I know you don't run the organization, but obviously you have a great deal of insight and influence and sit on the Board which oversees it. Ganesha811 (talk) 13:16, 20 March 2021 (UTC)[reply]
Just to provide some information on the point Nosebagbear raised about limits on this type of revenue: There are a couple of relevant legal requirements for public charities in the U.S. First, as a public charity, the Wikimedia Foundation must meet what is called “the public support test.” There are two ways to do this, but the least subjective is to have a minimum of one-third of our revenue come from donations from the general public. This has always been a very easy bar for us to clear with our reader donation model. There are similar limits on how much “unrelated business income” a public charity can receive. This is a fairly tedious area of U.S. tax law that I don’t dare try to explain here as a non-lawyer, but we can get you more information on this next week if there is interest.--Lgruwell-WMF (talk) 17:35, 20 March 2021 (UTC)[reply]

Amazon Web Services

I haven't seen anyone bring this up. You're using Amazon Web Services here? Not merely service as a software substitute, but run by Amazon itself. First you use Zoom for communication, and now this.

So here is my question: Do you stand with the free software movement and the free software concept that Wikipedia was built on any more, or are you leaving it behind because you consider it to be more profitable to forget about ethics?

You would be wise to heed the following:

A productive and ethical business can make money, but if it subordinates all else to profit, it is not likely to remain ethical.

— Richard Stallman

That's all I really have to say...

DesertPipeline (talk) 03:40, 20 March 2021 (UTC)[reply]

Dear DesertPipeline,
I should start by emphasising that all code that this team produces is, naturally, FLOSS. We have overtly stated that doing so is one of the core operating principles of the project. Currently, it is mirrored Git here.
With regards to the use of AWS as a hosting platform for the time being, I hope you will find documentation provided on the project's MediaWiki.org page useful – in the section titled #Application hosting. It states:

The engineering goal of this project is to rapidly prototype and build solutions that could scale to the needs of the Enterprise API's intended customers – high volume, high speed, commercial reusers. To do this, the product has been optimized for quick iteration, infrastructural separation from critical Wikimedia projects, and to utilize downstream Service Level Agreements (SLAs). To achieve these goals in the short term, we have built the Enterprise API upon a third-party cloud provider (specifically Amazon Web Services [AWS]). While there are many advantages of using external cloud for our use case, we acknowledge there are also fundamental tensions – given the culture and principles of how applications are built at the Foundation.

Consequently, the goal with the Enterprise API is to create an application that is "cloud-agnostic" and can be spun up on any provider's platform. We have taken reasonable steps to architect abstraction layers within our application to remove any overt dependencies on our current host, Amazon Web Services. This was also a pragmatic decision, due to the unclear nature of where this project will live long-term...

...We have intentionally kept our technical stack as general, open source, and lightweight as possible. There is a temptation to use a number of proprietary services that may provide easy solutions to hard problems (including EMR, DynamoDB, etc). However, we have restricted our reliance on Amazon services to what we can be found in most other cloud providers...

...We are looking to provide Service Level Agreements (SLA) to customers similar to those guaranteed by Amazon's EC2. We don't have equivalent uptime information from the Wikimedia Foundation's existing infrastructure. However, this is something we are exploring with mw:Wikimedia Site Reliability Engineering. Any alternative hosting in the future would require equivalent services or time to allow us to add more staff to our team in order to give us confidence to handle the SLA we are promising. In the meantime, we are researching alternatives to AWS (and remain open to ideas that might fit our use case) when this project is more established and we are confident in knowing what the infrastructure needs are in reality.

[In between the "..." in those quotes are lists of the specific pieces of technology used.]
With regards to Zoom: As I noted in both the my message on Wikitech-l announcing the two 'office hours' calls that were held yesterday; and the conversation on the WikimediaGeneral telegram channel about the video recordings of those calls being prepared for upload to WikimediaCommons - Other “office hours” meetings can be arranged on-request on a technical platform of your choosing. The team will come to wherever there are people interested to talk to us. For example, we'll be at the SWAN calls tomorrow. The community that organises those calls chose to host them on Google Meet. The reason I selected to Zoom as the platform those first two calls - which I had to select before knowing how much general interest there would be in attendance at them - is directly informed by the Lessons learned # Tools report from the Movement Strategy process.
Finally, with regards to our movement's interaction of money: at the very beginning of the positioning Essay I wrote for this project, I spend some time discussing the difference between 'gratis' and 'libre' in the concept of free as it applies to our movement's explicit choice to not have a non-commercial licensing restruction. It is also in reference to a quote from Ricard Stallman.
I hope this response goes some way towards addressing your concerns.
Sincerely, LWyatt (WMF) (talk) 10:37, 20 March 2021 (UTC)[reply]
User:LWyatt (WMF): I can't say I'm a good communicator, so I apologise for that message's abruptness, and this one's. I'm not exactly sure what to say with regard to your response, but I note that you say "FLOSS" and a quoted sentence contains a mention of merely "open source". May I ask why the Wikimedia Foundation is not committed to calling libre software what it is, rather than using terms like those? We must remember that freedom is the most important aspect, which is why it should be emphasised in the name. Regardless of whether or not the software is libre, though, I just wonder why Amazon Web Services has been chosen. I'm sure this is a purely practical decision, so would you consider thinking about the ethics here? All of us have an obligation to be ethical, especially to offset the damage that those who refuse to be ethical do. We must always think about the long-term consequences of our actions, a lesson which still has not been learned in the wider world, making it even more important. Regards, DesertPipeline (talk) 10:48, 20 March 2021 (UTC)[reply]
I am quite happy to change the teminology being used on the documentation if you think it is worthy of doing so. To clarify: are you suggesting that it would like me to change, on the MediaWiki page, the sentence "We have intentionally kept our technical stack as general, open source, and lightweight as possible." to say "We have intentionally kept our technical stack as general, libre and open source, and lightweight as possible."?
Yes the resaon that the service is currently being build and hosted on AWS is purely practical and pragmatic, and is being designed to be able to be moved as easily as possible. At the very end of that section of documentation on MediaWiki it states:

In the meantime, we are researching alternatives to AWS (and remain open to ideas that might fit our use case) when this project is more established and we are confident in knowing what the infrastructure needs are in reality.

So, if you have practical ideas for what we should do instead once the infrastructure needs are known, I recommend you share them on the talkpage on MediaWiki.
With regards to the question of ethics: One of the ethical considerations in using third-party hosting for this third-party-specific project, is the ethics of using donor's money wisely. If we were to build this infrastructure in-house from the start, it would require radically re-architecting the WMF systems in order to be able to serve the contractual obligations of Service Level Agreements with big tech companies. That is expensive in terms of time, hardware resources, staff, and computing power - which all equals expensive in terms of money. That would, in effect, mean subsidising the engineering requirements of the world's largest companies, with donor money. An unethical position in my opinion. Rather, by setting this service up on an independent platform, the existing resources (staff, infrastructure) of the WMF is undisturbed - and can continue serving the needs of editors and readers. LWyatt (WMF) (talk) 11:26, 20 March 2021 (UTC)[reply]
User:LWyatt (WMF): I appreciate your responsiveness. Also, do you consider it necessary to use the term "open source"? I think it would be better if it were just described as either "libre" or "free (libre)" – including the wikilink in either case as in my example here so that anyone not aware of the meaning can find out. Regards, DesertPipeline (talk) 12:23, 20 March 2021 (UTC)[reply]
Debates about linguistic preferences in the terminology is exactly why cumbersome acronyms like "F/LOSS" exist - in order to ensure the concept is all-inclusive. I wouldn't be surprised if we added the word free, and removed the word open, that someone else would be concerned for the equal-and-opposite reason. Suffice to say, whatever terminology the documentation uses for software we build - its intended meaning is as inclusive of the meanings of free-libre (and gratis) and open as is practicable, and for anything we build upon of others which is not wholly both free and open, then it will be as agnostic as possible - as per the blockquote above. LWyatt (WMF) (talk) 12:40, 20 March 2021 (UTC)[reply]
User:LWyatt (WMF): I consider the distinction important, because "free" is a lot more desirable than merely "open". It's true that libre software is technically also "open source" software, but the two terms are very different – where libre software is about pointing out the actual issues that need to be solved (a lack of freedom in computing), "open source" is merely about the practical benefits. This is why I think it's better to not use the term "open source" at all; it's redundant, anyway, because that which is free as in freedom must necessarily be "open" – either using the definition most would think ("you can view the code") or the one it is supposed to mean ("you have access to the code"). Because of these reasons, I consider there to be no reason to use it, even if "free (libre)" is written along with it (especially because that can still cause confusion and make people think of gratis rather than libre). Perhaps some will complain if you don't use that term, but if it is explained to them that the Wikimedia Foundation is committed to freedom, and has no desire to undermine that concept with terms that distract from that goal in favour of mere practical considerations, they will hopefully realise the error of their ways. I suppose that's quite an idealistic outlook, but I hope you see where I'm coming from. Regards, DesertPipeline (talk) 12:52, 20 March 2021 (UTC)[reply]
I've made this edit, which as you say might be redundant legally-speaking, but for the avoidance of anyone's doubt. LWyatt (WMF) (talk) 13:35, 20 March 2021 (UTC)[reply]
User:LWyatt (WMF): As I said before, I'd prefer if the term "open source" was not used for the reasons previously outlined. Is the term necessary here in your opinion? DesertPipeline (talk) 13:48, 20 March 2021 (UTC)[reply]
DesertPipeline, Yes I believe it is necessary for two reasons: because "open source" is a commonly used term, so if someone goes looking (e.g. by text searching on the page) I want them to find it; and because if I were to replace the words open source with the word libre, it would be too easy for some people to misinterpret it as my having actively removed the words "open source" which would seem controversial. So, keeping them both is probably best described as a kind of legal doublet or professional pleonasm - the purpose of which is to serve as a redundancy check. If there's anywhere else in the texts that could benefit from this redundancy check, please advise. LWyatt (WMF) (talk) 14:06, 20 March 2021 (UTC)[reply]

I think there's a bunch of very legitimate concerns around using AWS for this project. Amazon has a lot of very questionable practices in how it treats its employees, and the Foundation is indirectly supporting Amazon by using AWS. Arguably other major cloud providers (Microsoft Azure, Google Cloud Platform, etc.) also have their ethical downsides. And, although generally public cloud solutions such as AWS are actually excellent from a security and privacy standpoint, we really do want to avoid using non-Wikimedia infrastructure so that we have full control over those factors ourselves. The best thing to do, really, is to use Wikimedia infrastructure. In the short term however, I don't think it's quite that clear cut. I was going to write a long list of reasons why, but User:LWyatt (WMF) covered them already; cost efficiency, speed of prototyping, organisational overhead and project management etc. The biggest reason, to me, why using a public cloud system to prototype makes sense is to do with scalability. It's not clear right now how much this enterprise API will be used; we know there's interest in it from consumers, but will that interest actually materialise into use? Without knowing how much it'll be used, it's difficult to know how much and what kind of hardware to provision, which could either slow the project down or needlessly increase the cost. Additionally, there's only so much you can do to estimate the performance of a system without actually starting to prototype it, so that adds yet another variable to the uncertainty of what kind of hardware to provision. There's the potential to waste a lot of money on unnecessary hardware, and time (and therefore more money) on sizing, provisioning, racking, etc. all that hardware. Or, insufficient hardware could be procured, which then results in a scramble to find ways to deal with that problem, wasting more time and money. Public cloud solutions handle scaling out of the box, and they can actually be surprisingly cost effective in cases where you don't know in advance what kind of scale you'll need to operate at. Therefore I'm not concerned about prototyping this in AWS, but in the long run I'd want to see it migrate over to Wikimedia infrastructure once the prototyping has given a clearer picture of what the infrastructure requirements would be so that a plan can be made on how Wikimedia infrastructure can be modified and improved to support it. --Deskana (talk) 15:35, 20 March 2021 (UTC)[reply]

Thank you Deskana, well put. LWyatt (WMF) (talk) 15:48, 20 March 2021 (UTC)[reply]

This will effect donations negatively

If Wikimedia is getting a substantial amount of income from this proposed enterprise, the Foundation cannot in all seriousness claim that it will not have an effect on the large number of small donations -- currently the major revenue stream. The statement in the FAQ claiming otherwise is hand waving, pure & simple. A moment's thought will confirm this. Right now fundraising pitches follow the format "If you don't give us any money, we have no other revenue source to keep things running. We'll have to shut everything down." Further, it's reasonable to expect this enterprise will provide an income millions of dollars in size; otherwise, it will not justify the effort. Now even one million dollars is a lot of money to most people, so much it's hard to easily conceptualize. So when it is learned that Google, Bing, et alia are paying in total X million dollars, it will be a disincentive to users to donate. I know that if I had been making donations, & heard that this new revenue stream was bringing in millions of dollars, I'd be more inclined to spend my potential donation on beer than sending it in. Maybe this will not decrease the total of donations a large amount, but it will reduce it.

Now unless this potential loss in donations is properly estimated -- will it be a loss in the range of 2%? 5%? 25%? 50%? -- the reduction of total income will be a disorienting surprise. The first response will be to increase income from this revenue stream, which could lead to a vicious circle where keeping Wikipedia & the other projects afloat depends on selling a product to major corporations, who will -- intentionally or not -- force editorial changes. (For example, the fundraising department telling volunteers, either directly or indirectly, not to report some facts about Google because "they might not renew their contract". Salesmen always want to flatter their customers.) And if the community of volunteers perceive an outside group has an editorial influence, we will leave. Maybe slowly at first, but in the end too many volunteers will leave to keep the Wikimedia projects viable. At best, we will fork Wikipedia & other projects & abandon the Foundation; at worst we will drop out entirely from free knowledge. -- Llywrch (talk) 07:20, 20 March 2021 (UTC)[reply]

You've certainly outlined a worst case scenario quite well, but do you seriously imagine a situation in which the fundraising department imagines for even one moment that they can tell volunteer editors what to do? I can't imagine it and of course if they did it, a ton of volunteers would quite rightly go on strike immediately - me among them.
As to the question of whether a new revenue source might negatively impact negotiations and the need to estimate that - I do agree to an extent that it is possible, and I do think that the WMF has considered the issue carefully. However, to actually *estimate* it in a reliable way is so difficult as to be practically impossible. My personal belief is that it will have very very little impact. Here's why. For quite some time, the WMF has been managed well, financially, such that we bring in more money every year than we spend so that we can build up our reserves - which we have done. Additionally we have built up the WMF Endowment fund into something quite substantial. There are occasional news stories about this, basically saying "Why is Wikipedia asking for money, they have a ton of money already?" And the impact on donations has not been negative at all - indeed, I think it is arguable (and I know this in a direct way if we consider major donors who I've personally talked to) that having the WMF on sound financial footing, so that we can do more for free knowledge globally, is a stronger and more stable longterm incentive to donors, as opposed to pursuing what I would regard as folly: teetering forever on the edge of bankruptcy in order to panic people into donating money. That would be terrible!
There's another way of thinking about this as well. I know of a prominent water charity that gets major donations to cover their operating expenses in full, so that it can honestly promise smaller donors that their money will be 100% earmarked for work in the field getting clean water to people who need it. If it becomes possible to say "Hey, we get enough money from enterprise services to cover <this not very sexy part of our work, like bandwidth costs or such, whatever> so that 100% of your donation money goes to <supporting the growth of communities in the developing world, whatever> then that's likely to bring in more donation money as well. (I'm not arguing here whether those are the two things that would make sense - but the point is: if enterprise money pays for stuff donors don't care about, then more of their money can go to stuff that donors do care about).
Still - I happily concede and join you in an interest in closely keeping an eye on it. Suppose we found that for every $1 in revenue from enterprise resulted in a $2 loss of revenue from donations. The right answer would not be to increase income from the enterprise products, but to realize that it *costs money* to get that revenue. The right answer would be to cut the price to zero, as paradoxical as that might seem.--Jimbo Wales (talk) 10:16, 20 March 2021 (UTC)[reply]
I just sat down to write a reply here, refreshed the page, and found Jimmmy had beaten me to it! And with a much more detailed answer than I would probably have written! Thank you. Moreover, my answer would have focused a fair bit on the necessary oversight of the Board of Trustees to uphold and enforce their a) fiduciary responsibility to movement funds, and b) moral responsibility to movement principles. So, a board member replying directly, and quickly, and extensively, serves to highlight that oversight role is being fulfilled.
The only thing I would like to add is to refer back to Lgruwell-WMF's comment earlier on this page. She notes that the Donor Relations team, which answers hundreds of thousands of donor questions\inquires\complaints a year, typically receives many complaints from donors when there is news coverage of a stance the foundation is taking, or something new we are doing. The WMF Communications team has monitored 120 news stories in the last week, across many languages, about this topic. Yet, as Lisa states, the donor relations team have only received one donor complaint on this topic. This is, obviously, unscientific and is not meant to dismiss the need not be vigilant about it. Nonetheless, it is a good 'early-days' measurement of the sentiment among the large and diverse range of people who have, and continue, to donate. Rigorous monitoring of donor sentiment will continue.
In my personal monitoring of social media conversations about this project, the most common sentiment I am reading from the 'general public' is words to the effect of "its about time Big Tech paid their fair share" [I make no judgements about this opinion, just stating what I'm reading]. And the majority of Wikimedian commenters I've read are stating words to the effect of "this sounds fair enough but we should remain vigilant that it's done right" - which is a sentiment I think is entirely appropriate.
Sincerely, LWyatt (WMF) (talk) 11:09, 20 March 2021 (UTC)[reply]