How Well-Read Are Technical Wikipedia Articles?
Posted by Brian Kelly on 8 July 2010
In a recent post on Having An Impact Through Wikipedia I suggested that it would be useful if JISC-funded synthesis reports, for example reports on emerging new standards, used Wikipedia as a means of enhancing access to such work. In the post I pointed out that “I can’t find usage statistics for the page [but] I suspect that the article [on Amplified Conference which I created] will have been read my more people than have read my various peer-reviewed papers, blog posts, etc.” In response to a request for examples of tools which provide usage statistics for Wikipedia articles Martin Greaney suggested that “It’s quite basic, but the tool at http://stats.grok.se/ might give you enough of an idea of the traffic to certain articles in Wikipedia“.
As Lorcan Dempsey suggested in a tweet “The Wikipedia article traffic stats site mentioned in your comments is quite interesting. wonder how reliable is“. I agree and thought I would explore what the statistics tell us about Wikipedia entries for a number of areas related to Web, metadata and related standards of interest to the JISC development community.
My survey was carried out on 6 July 2010. The following table provides a link to the relevant Wikipedia article, the data the article was created (with a link to the original page for the article), my comments on the article and the usage statistics for October 2009 and June 2010 (two dates chosen to observe any significant variations).
|Page||Created||Summary (subjective comments)||Stats: Oct 2009||Stats: Jun 2010|
|Linked Data||May 2007||Multiple concerns have been identified with this article.||5,423||8,102|
|HTML||Jul 2001||Appears to be a well-written and comprehensive article. Includes info box so factual information is available in DBPedia.||147,357||143,386|
|XML||Sep 2001||Appears to be a well-written and comprehensive article. Includes info box so factual information is available in DBPedia.||159,749||126,599|
|XSLT||Feb/Jun 2002||Appears to be a very thorough and comprehensive article. Includes info box so factual information is available in DBPedia.||7,160||18,938|
|RSS||Sep 2002||Appears to be a well-written and comprehensive article. Includes a very brief info box so factual information is available in DBPedia.||7,160 (gaps)||18,938|
|AJAX (programming)||Mar 2005||Appears factually correct .||98,629||90,300|
|SAML (Security Assertion Markup Language)||Sep 2004||Appears to be a very thorough and comprehensive article.||4,421||2,942|
|Z39.50||Oct 2004||Brief article which has been flagged as in need of improvements.||3,960||2,592|
|Search/Retrieve Web Service||Feb 2004||Very little information provided.||506||462|
|Dublin Core||Oct 2001||Appears factually correct though citations need improving.||7,013||7,501|
|METS (Metadata Encoding and Transmission Standard)||Sep 2006||Appears factually correct though citations need improving.||7,236||4,573|
|MODS (Metadata Object Description Schema)||Aug 2006||Appears to be a reasonable although succinct summary.||546||583|
Extrapolating from the usage statistics for the two dates it would seem that popular articles such as HTML and XML have an annual number of views of around 1,750,000 and 1,720,000 whilst an article on a less well-known standard such as METS has an annual number of views of around 70,0000. It is perhaps surprising, in light of the high viewing figures for METS that the annual viewing figures for MODS is around 6,700. Perhaps this is due to the name clash between the METS acronym and the Mets name used to refer to the New York Mets. However there isn’t, as far as I am aware, such scope for confusion with names such as HTML, XML, SAML, etc.
What conclusion might we draw from such statistics? I would suggest that if I had an interest in ensuring that users had a good understanding of what Dublin Core is about and had access to the key sources of information then contributing to the Dublin Core Wikipedia page would be a good way of achieving that goal – after all the estimated viewing figures of around 87,000 surely can’t be ignored.
Now Matt Jukes pointed out the potential difficulties of getting content into Wikipedia. But that is a question of ‘How we go about contributing to Wikipedia?‘ rather than ‘Should we?‘
Can we accept that the answer to the second question should be ‘Yes‘ so that we can explore ways of addressing the first question?