Why does anyone bother contributing to Wikipedia?
by Tim Harford
in the Financial Times

I was very eager to see how a reputable news outlet like Financial Times would handle this interesting question. Unfortunately, I found Tim Harford's logic so puzzling, I was utterly disappointed in this pointless article.

Sounds like an interesting study though. His conclusion seems to be that wikis need critical mass or no one will contribute, because the contributors are just basically showboating and won't do it for no audience.

Agree, although his point is not entirely clear.

I agree. Citizendium is a classic example. After all the hype and media attention had passed - fewer people are contributing and the number of articles there have slowed to barely a trickle. Wikipedia has lasted for as long as it has because in its early years it was operating in an environment with few online competitors, and has been the centre of attention for a long time.

Yes, the conclusion is that people contribute to social media sites because they want attention. This is the motivation of a significant fraction of Wikipedia's editor base, probably either first or second most common, along with the "so that the truth can be known" motivation of those forwarding their personal ideologies.

Critical mass drives both of these motivations: people who want attention aren't going to go somewhere to get it if there's nobody there to notice, and people seeking to spread the truth gain nothing from shouting in an empty room.

Wanting attention isn't too bad a motive. When I'm asked to speak at a conference, I usually inquire as to the numbers and "quality" of the participants. I'm not so keen to speak if there answer is "not many and not much". OTOH, if I were asked to present to a room of professors I'm hardly likely to refuse.

Why on earth would anyone want to contribute to anything which has no audience and no feedback?

I wonder though that Wikipedia has a real problem. A lot of people contribute reasonable stuff, or do useful things, for a social networking payback. Five years ago, there was nothing (except specialist sites) to compete with Wikipedia for social networking among the educated and post-teenage user. Now most of these people are on Facebook - which has a much more attractive modern interface (and the resources to develop it) - while the wiki looks tired.

I fear that, over time, Wikipedia will lose out as social networking interfaces evolve and Wikipedia is stuck with low resourced devs and a very conservative community and so will not keep up.

The other problem is that as the userbase slowly decreases, Wikipedia has the dilemma of either seeing articles deteriorate, or locking more and more of them, since a decreasing userbase can't maintain a still increasing database. If they lock more articles (as they are doing) this will hasten the decline. It isn't so attractive to become a wikipedian if high-profile articles are locked, and others are seen as so "finished" that your contributions are reverted.

Note the fact that there are moves to semi-protect all Feature Articles, and articles like "Climate Change" have been semi-protected for months. This would not have happened two years ago.

Citizendium (and many others) is held back by a truly bizarre registration process, which would drive away anyone but the True Faithful. Yes, Wikipedia probably goes too far towards anonymity, but getting a Citizendium account takes about as long as getting a bank account, and the questions are more cryptic.

Idealists contribute because they hope someone will come to them and listen. Marketers don't bother unless there are large numbers of listeners because it isn't profitable.



Empires rise, then shine, and then fall, but don't disappear completely most of the time; in the online world it takes just about 2-3 years to see this cycle rather than 200-300 or more.

False distinction. Idealists look for a congregation to preach to. They don't bother speaking to the air.

I guess one of the problems about deposing the Wikipedia empire is that you need to attract a suitable population of serfs to do the grunt-work. The odd thing about Wikipedia is that the serfs don't actually see anything wrong with their current environment and are very defensive of it. You'd think with the obvious flaws in the way Wikipedia works that someone would be able to come along and fork it.

In a way, the blind, uncritical loyalty to Wikipedia, rather than support for the general concept of a worthy, user-contributed encyclopedia seems to be the biggest barrier to the immediate fall of the WMF empire. It is not surprising that WMF are so protective of the whims of its minions.

Surely that is as it should be. There come a point when all that can be done is polish the knobs, and the more that is done the more likely it is that you'll rub the gilding off.

Just curious, but has the ability to fork Wikipedia been demonstrated recently? Are the database dumps and such in order? Is the MediaWiki software not too corrupted?

My estimate is that it would cost hundreds of manhours and require some very specialized knowledge to get a fork of Wikipedia up and running. And I'm not altogether sure that a current database dump is even available or possible to get anymore.

I have an idea for the $11 million. Hire a full time technician whose job is to uphold the Foundation's promise of making the information easy to share. You'll still have $10.9 million. But I guess it is pretty low on the list of priorities.

There are websites that fork wikipedia but unfortunately most of them serve as content for porn, Viagra, acai berries, spam websites etc. If you're a serious researcher why would you go there? There is also the problem with Google - they tend to lower rank sites with duplicate content - so wikipedia always stays at the top of the tree.

I'm no expert, but I think a fork is something different from a scraper or mirror. I'm not talking about having some automated program including content on a site. I mean taking the articles and contribution histories, putting them into a MediaWiki installation, and then having human volunteers work on them wiki-style.

Technically getting the wiki source code through and functioning so that you can change it is a lot more challenging than merely getting the finished product through.

I don't know about others' experience, and I'd classify myself as being awfully clumsy at using Mediawiki tools. Nonetheless, I find the process of being able to copy entire categories of pages from Wikipedia over to Wikipedia Review to be almost frighteningly simple, using the WP Export tool followed by the MWB Import Pages tool.

Anyone is welcome to start building a human volunteer network to do just that, on my wiki. Come one, come all. Except Cyclopia.

Cyclopia? People with congenital abnormalities?

No offense, but these articles on Wikipedia Review tend to look crappy and lack contribution history. Example: http://www.wikipediareview.com/Nissan_Sentra

Someone at WMF should make the export/import functions work so that they seamlessly drag the entire contribution history and the pictures without a problem. Wikipedians should also avoid making every article so template-heavy that it needs ten templates dragged along with it wherever it goes or it breaks.

No offense taken. Point of fact, I believe that the "entire contribution history" is a check-box that you can select on the Export function. Pictures and templates are a pain in the ass, agreed. However, the core content really is the text, as that's what the search engines are primarily able to read and are looking for. And the text is what's in most dire need of fixing by more competent adults.

There are idealists of both stripes of course. Extroverts and introverts, I guess.

It's like the introvert accountant who looks at his shoes when talking to you. The extroverted accountant looks at YOUR shoes.

It's a box you can uncheck. Leaving it checked only gives you the most recent revision. Unchecking it will cause the process to fail, unless the article has a reasonably short history.

Five years ago yes, astonishingly simple. In 2010, it's just really crude and buggy, and it doesn't need to be that way anymore.

Agreed the text needs work too. Which is why I'm not so interested in a direct fork anymore.

edit - removed license speculation I don't want to get into

http://en.wikipedia.org/wiki/Wikipedia:Database_download

Here are the data dumps if anyone's interested.

I note that the full dump of articles appears to have successfully completed last January - 250GB worth. The latest revision dump seems fairly reliable.

That would suggest that it is not likely that an up to date fork containing a full revision history would be feasible, though a current scrape would be available. Is a current scrape sufficient for attribution purposes for a standalone project with no linking back to Wikipedia.

In practice, it seems simply giving credit to "Wikipedia" is enough to ward off any sort of licensing claim that anyone might be able to bring against you.

Of course, this was not a sufficient part of the terms of the GFDL, nor is it even sufficient for the migrated CC-by-SA license. But, as I said, what happens "in practice" is what probably matters.

In other news, about 90% of motorists habitually exceed the posted speed limit on roads.

Can I put "Banned from Wikipedia Review" on my signature?

Only if you include a "dofollow" hyperlink back to Wikipedia Review.com.

If I can link today's revision of Sir Isaac Newton, it's a deal.