Campaign for video, costs of crowdsourcing, iPhone app, brief headlines

Last week, the Open Video Alliance, an organization promoting "open standards, open source, and open content" for video, launched a campaign entitled "Let’s Get Video on Wikipedia", involving a portal site at VideoOnWikipedia.org. It is a collaboration with Mozilla Drumbeat, the Participatory Culture Foundation (makers of the Miro video platform)[1], and Wikimedia New York. Erik Möller welcomed the initiative on behalf of the Wikimedia Foundation, commenting:

We don’t expect that Wikipedia will turn into “Wikitube” anytime soon, but we do hope that thousands more relevant educational videos will find their way into articles in our projects.

As TheDJ explained, the portal uses the new HTML5 video player and other video tools developed by Kaltura's Michael Dale for MediaWiki.

On the English Wikipedia, the initiative is accompanied by the formation of the WikiProject Lights Camera Wiki!, and the development of guidelines at Wikipedia:Videos (in addition to existing pages about file uploading and usage).

In his article "Shirky and Sanger, or the costs of crowdsourcing" (appearing this month in the Journal of Science Communication), researcher Mathieu O'Neil examined the role of traditional notions of expertise in collaborative online knowledge production ("crowdsourcing"), comparing the differing approaches of Wikipedia and Citizendium (the online encyclopedia project founded in 2006 by Larry Sanger, Wikipedia's former chief organizer).

The article starts out by comparing Web 2.0 processes to the decision-making of the Internet Engineering Task Force and the free software movement, which eschewed authority derived from traditional hierarchies in favor of recognition of "autonomous technical excellence", as exemplified by David D. Clark's famous slogan: "We reject kings, presidents and voting. We believe in rough consensus and running code". While noting similarities to this culture in Web 2.0, which is still informed by the "hacker ethic", O'Neil identifies the fact that in today's non-hacker online collaboration "qualifications of participants are not always easily discernible" as an important difference.

As an example of such problems on Wikipedia, O'Neil highlights the case of William Connolley, who as a climatologist ran into trouble with an anonymous opponent while attempting "to correct mistakes on Wikipedia’s climate change article". According to O'Neil, the case "resonated deeply as it highlighted what can befall respected experts who wade into controversial wiki-waters" (it has also been mentioned in a 2005 Nature article and a 2006 New Yorker feature about Wikipedia), exemplifying the concerns of supporters of traditional encyclopedic approaches such as Britannica's and "the kind of incident which Sanger warned against in the early days of Wikipedia".

Citizendium tries to avoid such problems and reconcile traditional expertise with open online collaboration by requiring contributors to defer to "editors", users that have identified themselves as subject experts, for example by providing resumés. In an exchange with Sanger in fall 2006 [2][3][4], shortly after Citizendium had been proposed, Clay Shirky criticized this approach for its reliance on personal authority ("Deference, on Citizendium will be for people, not contributions ... Deference, on Wikipedia, is for contributions, not people...") and argued that Citizendium's approach was doomed to fail because of the costs (efforts) which are implicit in certifying expertise, deferring to it and policing to ensure such deference:

"If users do not want to participate in a system where the costs of participating are high, Citizendium will simply fail to grow."

O'Neil takes this 2006 debate as a starting point for his evaluations of crowdsourcing, interpreting Citizendium's lack of growth as evidence for Shirky's position:

"Several years later, it appears, in the first instance, that Shirky's criticism was well-founded: while Citizendium articles are of reasonably good quality, they are not very numerous. New participants to Wikipedia know that their contributions will have a significant audience; becoming a Wikipedia editor is trivial and instantaneous; since it lacks this immediate quality, Citizendium failed to attract the crowd."

However, O'Neil also lists numerous well-known concerns about Wikipedia's approach, identifying them as direct and indirect costs of "anonymous crowdsourcing". As direct costs (defined as "costs directly affecting the quality of the product"), he names:

• Uncertainty - while Wikipedia's approach can generate excellent results, the reader cannot be certain that it did so for a particular article; the quality is not consistent. According to O'Neil, one reason for this is that evaluation of the work of peers (cf. Linus' law) does not work as well on Wikipedia as it does in programming, because hackers can immediately evaluate the quality of their peers: "... either the code runs, or it doesn’t. In the case of Wikipedia, editors are not necessarily equipped to assess the accuracy of contributed information ..."
• Lack of perspective - O'Neil gives an example of an article about an important topic being shorter than one on a trivial topic (cf. Wikigroaning).
• Irresponsibility - O'Neil criticizes an alleged WP:SOFIXIT attitude towards the victims of BLP cases such as the Seigenthaler affair, again highlighting a difference to the world of free software development: "Passing off the responsibility for dealing with the failings of software or information onto the user may be an adequate response when people choose to participate in a venture, as is the case with free software projects. It is absurd when people have no say as to whether they are being written about."

On the other hand, O'Neil questions "whether there really is no deference to traditional expertise on Wikipedia" as assumed by Shirky, citing a 2009 paper by Larry Sanger (summarized in an earlier Signpost issue) where Sanger pointed to Wikipedia's reliable sources guideline and claimed that in several areas on Wikipedia, deference to expertise exists in practice. O'Neil summarizes: "Technical experts who create quality content command the respect of their interlocutors, and it is only when non-technical topics are discussed that deference breaks down, edit wars erupt and manipulation becomes possible."

As the indirect cost of Wikipedia's model (defining indirect cost as those which "divert resources from the task of building an encyclopaedia"), O'Neil names "many-to-many fighting", exacerbated by conflict of interest editing for public relations purposes:

"Shirky's calculation of costs and benefits does not take into account this unavoidable consequence of a situation in which scientists, interested amateurs, consumers, advertising agencies, and industry spokespersons come together to debate the proper definition of reality with no clear means of telling who is an expert."

O'Neil goes on to mention the Arbitration Committee, Newbie biting, article ownership attitudes, wikilawyering and sockpuppet abuse, and cites studies pointing to a possible bias against casual editors.

(See also last year's Signpost review of O'Neil's book Cyberchiefs: Autonomy and Authority in Online Tribes and O'Neil's responses)

Ars Technica recently reviewed Articles for iPhone, a new Wikipedia client for the iPhone and iPod Touch from Sophiestication Software. (Several other Wikipedia clients are already available for the iPhone, both free and fee-based.) According to the review, "[t]he developer cited the lack of a solution with attractive article layout, including the Wikipedia mobile site, as a major reason as to why they pursued this project". Ars Technica described the collapsing of infoboxes as the app's main difference from the standard view on mobile Safari, and also highlights a "Nearby" feature as a possible justification for the $2.99 price (a map with links to geotagged Wikipedia articles in the vicinity, which is offered by other iPhone clients, too).

(Earlier Signpost coverage about Wikipedia on the iPhone: March 2010, October 2009, January 2009, December 2008, February 2008)

• For the first time in almost two years, Wikipedia is mentioned on the front page of an Australian newspaper. The Australian Financial Review's lead story on Tuesday, titled "Labor under fire on broadband", claimed that a confidential CSIRO report on the National Broadband Network cited Wikipedia three times, including two technical citations comparing the performance of satellite and HFC broadband networks. The report has been dismissed by independent Senators Nick Xenophon and Steve Fielding as a "pamphlet".
• Jimmy Wales praised Google's recent decision to stop censoring search results in China, as mentioned by several media reports (which also referred to the past blocking of Wikipedia in mainland China). With these remarks Wales was reinforcing comments he made last week in his keynote at The Guardian's Changing Media Summit 2010.
• Reason magazine's Greg Beato asks "Where's Wikipedia" in a New York University list of the top ten works of journalism of the decade? The editorial goes on to note that only 3 of the 80 contenders were originally published on the web and praises Wikipedia for making "old news more accessible".

← Previous "In the news"

Next "In the news" →

In this issue

22 March 2010 (all comments)

Wikipedia-Books

News and notes

In the news

WikiProject report

Features and admins

Arbitration report

+ Add a comment

Discuss this story

These comments are automatically transcluded from this article's talk page. To follow comments, add the page to your watchlist. If your comment has not appeared here, you can try purging the cache.

Video campaign

• I'm unenthusiastic about the video campaign. I expect we'll mostly get videos of very little use added to articles with the way the campaign is going. I also think that there is really a limit to the number of videos that can sensibly go in one article (but there certainly is no limit to the number of videos that can go on Wikimedia Commons). —innotata 20:55, 24 March 2010 (UTC)[reply]

This article on the Chronicle of Higher Education's blog was unfortunately published too late to be included in the Signpost story, but it contains some interesting comments by the general director of the Open Video Alliance (about ongoing talks with universities). Regards, HaeB (talk) 16:13, 25 March 2010 (UTC)[reply]

Seems even more dubious. Watermarked low-resolution videos of lectures? I hope not. —innotata 15:11, 26 March 2010 (UTC)[reply]

Citizendium and Shirky

This week I somehow ended up as the main Signpost reporter for the "News and notes"and "In the news" sections, including the coverage of O'Neil's article. In that capacity, I'll refrain from adding my own observations to that summary, or turning it into a full-fledged review (last year Ragesoss wrote one of O'Neil's book, but as he is acknowledged in this paper for comments, I assume he won't do so now).

But having given a talk on Citizendium at last year's Wikimania which touched many of the same topics (including an interpretation of CZ's disappointing statistics as vindication of Shirky's 2006 criticism), I can't resist adding some remarks here on the talk page:

• Perhaps surprisingly, several independent commentators (Connolley among them) have voiced concern that Citizendium might actually be more hostile to scientific experts in some cases, in favor of what critics would call pseudoscience or fringe science. The folks at Rationalwiki offer a scathing criticism of Citizendium as a "crank magnet". Topics where Citizendium's coverage has been subjected to such criticism include homeopathy, memory of water, Intelligent Design, Young Earth Creationism, global warming and chiropractic. Conversely, Citizendium has been praised by opponents of mainstream science views, such as homeopathy lobbyist Dana Ullman (who got banned for a year on Wikipedia, but managed to get his Citizendium article on homeopathy approved status), or the climate change sceptics at the right-wing Heartland Institute, which recommend the Citizendium article on Global Warming on their home page [5]. I believe that this seemingly counter-intuitive development merits further research. To me, one reason appears to be that some of Sanger's assumptions about experts (to exaggerate: debates among academics are always civil, only dumb laypersons ever disagree with an expert, anybody with a PhD has to be regarded an authority in his area) are too simplistic. The problem is exacerbated by the lack of participation and by civility policies encouraging "ownership" issues of the kind that O'Neil mentioned (in the case of Wikipedia) on p.4.

• Besides the stagnating or shrinking participation numbers, another telling metric which is not obvious from the diagrams at citizendium:CZ:Statistics is the shrinking median article size (that diagram ends July 2009, but the downward trend has since continued).

• "Inconsistency" (of article quality) can - and should - also be regarded as a cost of Citizendium's approach (after all, it is still wiki-based and open to everyone). The project readily admits that most of its articles are in an unfinished ("live") state. But one of its central tenets is that the small number of its "approved articles" do provide the kind of certainty that the reader is missing at Wikipedia:

"Our 121 expert-approved articles are reliable and of world-class quality, rivaling the best printed encyclopedias."

This is a dubious claim, as is evident from the list I compiled here: User:HaeB/Citizendium approved errors (many of them basic proofreading oversights which would certainly not be expected at Britannica, WP:EBE notwithstanding).

Independently, the OnWikipedia blog recently examined one Approved CZ article and found numerous grammatical errors and factual inaccuracies, too. (Their analysis of the reasons for the "failure of Citizendium" is worth reading, it agrees with some of O'Neil's conclusions and adds others.)

It seems quite clear that an important reason for such quality problems is the complicated bureaucratic process required to make changes to an already approved article, and the inability of outsiders to correct errors in any article (or even just notify CZ of them). And despite of its failure to catch the errors in these examples, the approval process itself seems cumbersome and time-consuming; the number of approved articles has stagnated at 121 since mid-December.

• O'Neil's article also does not describe the costs of CZ's real name policy, which in my observation are considerable. Not only the cost of enforcing it (after self-registration brought in fake identities and vandalism, CZ tightened the process and made it quite cumbersome), but especially the privacy concerns which are associated with having one's actions (even minuscule ones like a typo correction) publicly recorded, with precise time stamps, forever . For example, Wikiweise (a fork of the German Wikipedia bearing many similarities to CZ), disabled contribution lists for users, to balance their own real name disclosure requirement [6].

• The link in footnote 9 (http://en.wikipedia.org/wiki/Wikipedia:Requests_for_arbitration/Climate_change_dispute_-_Final_decision) is not working, Wikipedia:Requests_for_arbitration/Climate_change_dispute#Final_decision seems to be the intended location.

• The Seigenthaler case is perhaps not the best example for the "Irresponsibility" point, as the perpetrator in that case was actually identified and temporarily lost his job because of his actions on Wikipedia.

• But overall, O'Neil's article is certainly very well informed and factually accurate (even in comparison with other academic writing about Wikipedia). One thing I really liked was the insightful comparisons of mass online collaboration as exemplified by Wikipedia (to avoid the controversial "crowdsourcing" term) to hacker culture, including a careful look at the differences and the problems they create on Wikipedia. I don't know if these ideas have been formulated elsewhere, but the paper certainly deserves to be cited for them alone.

Regards, HaeB (talk) 21:47, 24 March 2010 (UTC)[reply]

I found both the article proper, and your commentary here, to be well-written and informative. The article proper appears well balanced and fair, and even your commentary makes it clear what your opinions are and why you hold them. I say good job! —DragonHawk (talk|hist) 00:55, 25 March 2010 (UTC)[reply]

Agreed. Thanks HaeB. Rd232 ^talk 12:19, 25 March 2010 (UTC)[reply]

The more I read of Citizendium, the more it seems like Nupedia. It also seems to be running into the same difficulties that project did so I am surprised to see no mention of Nupedia in this article. Rmhermen (talk) 19:19, 26 March 2010 (UTC)[reply]

How to earn $30,000 within one week by making Wikipedia content just a little more accessible

An interesting fact about the reviewed iPhone app: According to the company, "Within a short week, Articles [for iPhone] sold over 10,000 copies."

Regards, HaeB (talk) 05:43, 26 March 2010 (UTC)[reply]