SUMMER INSTITUTE IN NURSING INFORMATICS CALL FOR ABSTRACTS

The University of Maryland School of Nursing 16th Annual Summer Institute in Nursing Informatics Planning Committee is seeking high-quality abstracts for peer-reviewed paper and poster presentations. The theme of the July 2006 conference is Advancing Clinical Practice Through Nursing Informatics; $20 000 is available in awards and scholarships.

Deadline for abstract submission is March 1, 2006. For details, guidelines, and forms, please visit: http://www.nursing.umaryland.edu/informatics.

NATURE STUDY FINDS WIKIPEDIA NEARLY AS ACCURATE AS BRITANNICA

In December 2005, the Web news service of the journal Nature, [email protected], reported on a peer-reviewed comparison of online content from Wikipedia, the collaborative Web encyclopedia founded by Jimmy Wales, and Encyclopedia Britannica. Organized by Nature staff, the comparative review found that Wikipedia's science entries rivaled the quality of Britannica's for accuracy of information.

The staff selected 50 entries from Wikipedia and Britannica, and stripped each of formatting, reference lists, and any metadata that might identify the source, to produce a pair of blinded entries that were then reviewed for accuracy by experts in each field. Reviewers were asked to look for factual errors, critical omissions, and misleading statements in the sets of entries, and 42 usable reviews were returned.

The news team then examined the reviews, evaluating the critiques from the standpoint of "typical encyclopedia users," which resulted in discarding some identified errors, such as omissions that experts had cited as critical but were not seen as essential for a general understanding of the topic. Some of the errors identified were judged simply to have been a result of poor phrasing, so they were also ignored unless they significantly hindered understanding.

Although numerous errors in both sets of entries were found, the difference in accuracy was not as great as expected. Wikipedia entries contained an average of four errors each, while Britannica's contained three. Only eight serious errors (eg, misinterpretations of concepts) were found in the entire test set, with both encyclopedias containing four of these inaccuracies. Overall, reviewers found 162 factual errors, omissions, or misleading statements in the Wikipedia entries, and 123 in Britannica.

Wikipedia, founded in 2001, comprises almost 4 million articles in 200 languages and, according to the Web-ranking service Alexa, is in the top 40 most visited Web sites on the Internet. It has encountered some high-profile problems; for example, a recent article suggested that one of Robert Kennedy's assistants may have been involved in his assassination, which was revealed to be false. Wikipedia has also been criticized for giving prominent place to controversial scientific theories because it accepts entries and allows editing by any Internet user, but the speed with which content can be updated and critiqued has also been pointed to as one of its strengths.

Nature reviewers' most-issued complaint about Wikipedia entries was the poor quality of articles, from misused language to confusing structure. Wales seeks to improve Wikipedia by enticing researchers to contribute works and edit existing entries, and plans to introduce "stable" versions of entries once an article reaches a specific threshold of quality. "Live" versions still undergoing editing will be rated by users in a trial version of an evaluation system sometime in 2006.

To read the Nature news item and download supplemental material on the study, visit http://www.nature.com/news/2005/051212/full/438900a.html.

WEBCITE CONSORTIUM OFFERS "PERMANENT REFERENCES"-ARCHIVING OF CITED ONLINE MATERIAL

Web pages or Web sites cited in scholarly articles are often important references worth preserving for readers, yet because of the volatility of the World Wide Web, many simply disappear and are no longer available as a reference for future scholars.

WebCite, an archiving service initiated by Gunther Eysenbach, MD, Editor and Publisher of the Journal for Medical Internet Research and an Associate Professor at the University of Toronto, can take a snapshot of a cited Web page and will store a copy of the HTML, including images or other files such as Adobe Acrobat documents, on its server. It is a tool specifically designed to be used by authors, readers, editors, and publishers of scholarly material, allowing them to permanently archive cited Internet references for scholars reading the cited article 1, 3, 5, or 10 years from now.

The caching (archiving) process can be initiated by authors prospectively (before publication) or by authors, editors, copyeditors, or publishers when an article is written, edited, or published. WebCite also has a crawler capability, allowing limited retrospective (after publication) archiving, although there is no guarantee that pages will still be available and that what is being archived is what the author saw when the online document was first cited.

WebCite offers more than systems such as digital object identifier and CrossRef, which ensure the stability of cross-links to other journal articles and other materials carrying a digital object identifier, but cannot ensure that cited nonjournal Web sites without a digital object identifier (eg, links to questionnaires on the Web, research reports, or quotes from home pages) will remain accessible to readers.

Continuous archiving by services such as the Internet Archive (Wayback Machine) or Google is performed randomly by a crawler, which does not focus on academic references. The caching process cannot be initiated by authors, editors, or publishers who wish to preserve a specific online reference as it was when it was quoted or cited.

Caching and archiving Web pages is not considered copyright infringement as long as the copyright owner has the ability to remove the archived material and to opt out. WebCite honors robot exclusion standards and no-cache and no-archive tags. Copyright owners of archived Web pages who wish to opt out of the database can contact WebCite staff to have the material removed.

WebCite charges article authors no fee. Although there are various possible models to cover the ongoing costs of operations, it is hoped that publishers will pay a membership fee (similar to PILA/CrossRef membership fees) to have cited Web references archived. Readers from publishers or journals who are WebCite members will also have free access to archived material, unless publishers opt to charge their readers or to make this a value-added service for subscribers only.

For more information, or to become a member of the WebCite Consortium, visit the site at http://www.webcitation.org, or contact Dr Eysenbach at [email protected].

REFERENCE