返回资讯中心
外部精选
软件工程
#开发者实时流

维基媒体基金会公布支出情况

来源标题:Wikimedia Foundation Spending

维基媒体基金会公布了其支出情况,资金主要用于运营、技术基础设施和项目支持。该基金会依靠捐赠运作,支出结构体现其对平台维护和内容生态的投入。

Hacker News 最新germanrabbit2026/11/7 · 9 天前30 分钟阅读

以下正文同步自 Hacker News 最新,版权归原站所有,已转换为易读排版。

In biology, the hallmarks of an aggressive cancer include limitless and exponential multiplication of ordinarily beneficial cells, even when the body signals that further multiplication is no longer needed. The Wikipedia page on the wheat and chessboard problem explains that nothing can keep growing exponentially forever. In biology, the unwanted growth usually terminates with the death of the host. Exponential spending increases can often lead to the same undesirable result in organizations.

Consider the following example of runaway spending growth:

In 2005, Wikipedia co-founder and Wikimedia Foundation founder Jimmy Wales told a TED audience:

So, we're doing around 1.4 billion page views monthly. So, it's really gotten to be a huge thing. And everything is managed by the volunteers and the total monthly cost for our bandwidth is about US$5,000, and that's essentially our main cost. We could actually do without the employee … We actually hired Brion because he was working part-time for two years and full-time at Wikipedia so we actually hired him so he could get a life and go to the movies sometimes.

According to the WMF, Wikipedia (in all language editions) now receives 16 billion page views per month. The WMF spends roughly US$2 million a year on Internet hosting and employs some 300 staff. The modern Wikipedia hosts 11–12 times as many pages as it did in 2005, but the WMF is spending 33 times as much on hosting, has about 300 times as many employees, and is spending 1,250 times as much overall. WMF's spending has gone up by 85% over the past three years.

Sounds a lot like cancer, doesn't it? For those readers who were around three years ago, did you notice at the time any unmet needs that would have caused you to conclude that the WMF needed to increase spending by $30 million dollars? I certainly didn't.

From 2005 to 2015, annual inflation in the US was between 1% and 3% per year, and cumulative inflation for the entire decade was 21.4%—far less than the increase in WMF spending. We are even metastasizing the cancer by bankrolling local chapters, rewarding them for finding new ways to spend money.

Nothing can grow forever. Sooner or later, something is going to happen that causes the donations to decline instead of increase. It could be a scandal (real or perceived). It could be the WMF taking a political position that offends many donors. Or it could be a recession, leaving people with less money to give. Whatever the reason is, it will happen. It would be naïve to think that the WMF, which up to this point has never seriously considered any sort of spending limits, will suddenly discover fiscal prudence when the revenues start to decline. It is far more likely that the WMF will not react to a drop in donations by decreasing spending, but instead will ramp up fund-raising efforts while burning through our reserves and our endowment.

Although this op-ed focuses on spending, not fundraising, it could be argued that the ever-increasing spending is a direct cause of the kind of fund-raising that has generated a storm of criticism. These complaints have been around for years, leading one member of a major Wikimedia mailing list to automate his yearly complaint about the dishonesty he sees every year in our fundraising banners.

No organization can sustain this sort of spending on a long-term basis. We should have leveled off our spending years ago. Like cancer, WMF spending is growing at an exponential rate. Like cancer, this will kill the patient unless the growth is stopped.

The reason I have so little faith in the WMF's ability to adapt to declining revenues (note that I specified the WMF; I think Wikipedia has shown an excellent ability to adapt to multiple problems) is the horrific track record they have regarding adapting to other kinds of problems.

In particular, their poor handling of software development has been well known for many years. The answer to the WMF's problems with software development has been well known for decades and is extensively documented in books such as The Mythical Man-Month and Peopleware: Productive Projects and Teams, yet I have never seen any evidence that the WMF has been following standard software engineering principles that were well-known when Mythical Man-Month was first published in 1975. If they had, we would be seeing things like requirements documents and schedules with measurable milestones. This failure is almost certainly a systemic problem directly caused by top management, not by the developers doing the actual work.

After we burn through our reserves, it seems likely that the next step for the WMF will be going into debt to support continued runaway spending, followed by bankruptcy. At that point there are several large corporations (Google and Facebook come to mind) that will be more than happy to pay off the debts, take over the encyclopedia, fire the WMF staff, and start running Wikipedia as a profit-making platform. There are a lot of ways to monetize Wikipedia, all undesirable. The new owners could sell banner advertising, allow uneditable "sponsored articles" for those willing to pay for the privilege, or even sell information about editors and users.

If we want to avoid disaster, we need to start shrinking the cancer now, before it is too late. We should make spending transparent, publish a detailed account of what the money is being spent on and answer any reasonable questions asking for more details. We should freeze spending increases to no more than inflation plus a percent or two, build up our endowment, and restructure the endowment so that the WMF cannot dip into the principal when times get bad.

If we do these things now, in a few short years we could be in a position to do everything we are doing now, while living off of the endowment interest, and would have no need for further fundraising.

Nope, it's not cancer. It's exponential growth.
Exponential growth can have the effect of a cancer, but not necessarily so. Cf. Moore's Law, which predicts exponential growth and held for decades with no ill effects (not for end users anyway), and even when it stopped being valid it didn't disrupt the end user's experience.
There would be cancer if the growth disrupted the functioning of WP or WMF, but it does not - cancer kills by claiming space that other organs need, and possibly by claiming metabolic energy, but money in WMF does not do any of that. More money is merely an enabler, for positive as well as negative things, but that's all.
Not all cancers are exponential in growth either, so the analogy breaks down both ways.
That said, exponential growth does come with problems which do need to be addressed, so the concerns raised may be valid points. I have too little insight into the details; I am just objecting to using a misleading analogy, giving all the wrong memes and making people with not enough time to read all the details consider the wrong options, plus I am not too happy seeing that more time is invested in getting the data than in presenting it with the right analogies, it's a strong agenda-driven bias which is its own kind of misconduct, quite on par with the misconduct being attributed to WMF in this article.
Joachim Durchholz (talk) 07:33, 8 May 2017 (UTC)reply

Among the many things wrong with this piece is the insinuation (implicit in the cancer metaphor) that the growth in expenses was unplanned and uncontrolled. If one compares the five-year spending projection from the 2010-15 strategic plan, one finds that the actual development stayed pretty close to the plan, which had been developed in a collaborative process with extensive community involvement and lots of discussion about what to spend money on - and what not. And underspending compared to plan has actually repeatedly been an issue.

"I have never seen any evidence that the WMF has been following standard software engineering principles" - that might be because you did not bother to look. Quarterly goals are available for all engineering teams on mediawiki.org, and the product teams largely follow the principles of agile software development (in particular Scrum). Agile might not use the exact same elements as your favorite 1980s era formulaic software process specification, but it is pretty much the standard today. BTW, is it correct to assume that the consulting business which is mentioned in the article intro (as evidence for the author's alleged expertise about the topic) is not about software, but mainly about other forms of engineering?

At https://phabricator.wikimedia.org/ , it's possible to follow along the work in detail, minute-by-minute. I would invite everyone to take a look in detail at the tasks that are being worked on right now - pick one at random from the "Activity Feed" or "Fresh tasks" boxes, for example, try to understand what its purpose is, what the downsides of ignoring the task might be, and then form an opinion about whether or not WMF should spend money on it. (Or, about how many of them could actually be imposed on volunteer developers, as the article tries to make use believe is possible for almost any useful software work for Wikimedia sites.)

The discussion about how much money WMF should spend, and on what, is of course absolutely legitimate and necessary. But simply defining growth as inherently bad is a fallacy. Obviously any organization has to grow from zero to its desired budget size, so with it will always be possible to cry "OMG 1000% growth!!". And yes, with all due respect to Jimmy, I think that that talk twelve years was dangerously mistaken about what would be needed even for the minimal goal of keeping Wikipedia online, stable and safe, without ever improving the interface, add new features etc.

Besides that, the amount of technical debt Wikipedia's software has accrued during those many underfunded years is enormous, and still felt today. (Just as a small example: To handle the rising traffic load, developers back then had to resort to very aggressive caching strategies, which succeeded in keeping Wikipedia available to readers at a bearable speed, but had the side effect of imposing a performance penalty on registered users - for which pages can't be cached - and anyone who edits a page. I.e. editors, who are the lifeblood of the project, were precisely those for whom it was the most sluggish. This only began to be mitigated a little over two years ago, when enough resources became available for seriously starting to work on performance.) And Wikipedia still has much less paid tech staff than basically any other website of its size.

(Personal comments, informed both by my experience as WMF employee and as community member since 2003, who for many years prior to joining WMF observed it from the outside, including writing about its activities and spending as former editor-in-chief of the Signpost.)

Regards, HaeB (talk) 03:27, 28 February 2017 (UTC)reply

Re: "is it correct to assume that the consulting business which is mentioned in the article intro (as evidence for the author's alleged expertise about the topic) is not about software, but mainly about other forms of engineering?", that is correct. I do a lot of work with software, but on chips that typically cost less than five cents, have 256 nybbles of RAM, and are not programmable -- meaning that we write raw assembly language code with no hardware to run it on other than a few crude simulation tools, then order masked-ROM parts (minimum order 50,000 chips with 6 to 8 weeks from code submission to chip arrival) and throw them all away if there is a bug. Clearly my experience isn't even close to the same as what the WMF developers do, and thus my opinions on this are suspect. Getting back to your main point, I remain 100% convinced that the WMF developers are doing the right thing and are using techniques that we who are in the business of producing 100,000 electronic toys per hour know nothing about. Nonetheless,

正文由 探How 从来源站点 RSS 同步,内容未经改写;遇到排版缺失或需要图片、视频时请以原文为准。

打开原文
去社区讨论 →