SUBSCRIBE NOW!

Vol. 28 No. 5 — Sep/Oct 2004

On the Internet, the traditional information cycle is broken in a variety of ways. News may be reported, analyzed, debated, corrected, and reinterpreted in a matter of hours. Old stories from decades ago may be re-examined. Factual information can be evaluated, expanded upon, and expounded on by a wide variety of readers.

Instead of reading through complete Web pages or sites, searchers can browse results and choose to read a variety of extracts from pages created by completely different organizations. Finding a community of Web sites that together provide an answer can offer a deeper and broader understanding of certain issues.

THE WEB AS INFORMATION COMMUNITY

The Web has succeeded so spectacularly as a new publishing and communication medium for many reasons—the ease with which anyone can publish, the ability to change and update content, the interconnectedness from linking, the lack of a limit to the quantity of information published, and more. While many Web sites, including some of the most popular ones, continue to use the print model of publishing information in somewhat static articles, others are experimenting with improving overall information quality by having broader participation in the writing, correcting, and updating of content.

With the linking patterns on the Web, sites can create virtual communities of interlinked sites that provide different views, related information, and varying interpretations while still linking to each other. Following the links between the sites can create a more complete information portrait of an issue.

SINGLE SOURCE DOMINANCE

Still, for many online information seekers, a single source information focus remains. When an information need is of relatively low value, a single Web page will satisfy most users. Simply looking for the stars in the movie Rear Window, the meaning of "photosphere," or the five stages of grief? A Web search on any of these will pull up plenty of pages, all of which will probably have a correct answer. For those just looking for answers for their own curiosity, to help a friend, or on a whim, the single page can work.

For information professionals, there are times when an answer on a single page may suffice, but more often confirmation from several diverse sources helps confirm authenticity of the information. Yet with the Web, authenticity and accuracy is always questionable. Many pages, even from reliable organizations, have typographical errors and misstatements of fact. It is so easy to post a Web page that much Web content fails to have significant editorial oversight.

VARIABLE CONTENT

For example, in looking for an explanation of a biological process, a USGS (U.S. Geological Survey) Web page (from the biology side, not the geology side) gave one explanation that did not match the text of the search query. Checking the current page against older copies from the Wayback Machine showed that there had been a small change on the page—a "not" had been removed. This small removal completely changed the explanation of the process, and made it match the definition from other reputable sources. But it goes to show how even reputable, often authoritative organizations can make simple errors on Web pages.

Consider someone looking for the Spanish way of saying "happy birthday." Many Web pages contain the Spanish phrase, but different ones have variations such as

¡Felíz cumpleaõs!

Feliz Cumplean~os!

Feliz cumpleanos

The diacritic mark may be left out on some sites or may not display correctly.

In both these cases, trusting a single page, no matter how reliable the organization, can lead to incorrect information. Fortunately, the Web makes it easy to move beyond the single article focus.

BENEFITS OF MULTIPLE RESULTS

Search engines typically default to showing 10 results to a query, with Yahoo!'s default of 20 a welcome exception. Yet even with just 10 results, the results should be scanned to see how much difference they provide in their answers. Using an advanced search form, or the preferences to display more results, helps to further explore the possibility of conflicting or contradictory information.

Both of these examples showed a variety of conflicting answers in the results list. This led to the exploration of the contradictory or conflicting answers which when combined with evaluating the sources, comparing wording, and checking the frequency of the various answers, helped deduce the most likely correct answer to each.

THE BATHTUB QUESTION

The ability to triangulate on the Web and use multiple sources to come up with an answer is often much easier than it is to do in books and articles. Take for example the highly entertaining, if somewhat trivial, issue of when the first bathtub appeared in the White House. In the print era of book and article dominance, H. L. Mencken wrote an article for New York's Evening Mail in 1917. It discussed the "history" of the first bathtubs in America and the controversy around the installation of the first one at the White House by Millard Fillmore. The only problem is that this article by Mencken was fiction. After finding his "history" had been quoted as fact by other writers, Mencken wrote another article in 1926 in the Chicago Tribune as a public confession that his earlier piece was pure fiction and explained his reasons. Note that this took 8 years in the print age of the article.

By that point, his earlier fiction had been repeated so often that it continues to this day to appear in reputable reference sources, in print and online. Grolier's encyclopedias, the Washington Post, and the Internet Public Library [www.ipl.org/div/potus/mfillmore.html] have all taken information from that oft-repeated 1917 article and treated it as fact. For more information on this, see the book The Bathtub Hoax, and Other Blasts & Bravos from the Chicago Tribune by Mencken and the Web page, "Millard Fillmore's Bathtub" [www.sniggle.net/bathtub.php], which lists many places that have repeated the falsehoods.

Note the difference with how the Web can handle this kind of situation. Searching for white house bathtub at Yahoo!, Google, or Teoma finds a collection of Web pages including the Sniggle.net page and ones that credit Fillmore or even his successor Pierce for some reason. One of the best results for this question comes from a page that reprints a 1990 article from Plumbing and Mechanical on the history of plumbing in the White House. It discusses the hoax along with earlier reports of tubs in the White House. But for this question, no one single Web page really answers this question well. It is the sum total of the Web pages, incorrect and accurate, along with the reproduced articles that really help answer this question.

COMMENTS AND CORRECTIONS

It is the ability of the new online environment to quickly and easily correct, or at least criticize, information that makes the online medium so different from print. One problem with the published world of information as seen in books and periodicals is that despite editing, fact checking, and the peer review process, all kinds of errors still found their way into print, as the Mencken hoax illustrates. Periodicals would use errata sections to correct some of the errors, and letters to the editor could be used to debate a previous article's contentions and possibly set the record straight.

Unfortunately, many readers would never see errata and letters that, by the necessity of the printing process, would appear in subsequent issues of the periodical. While some indexes did a great job of combining both original article and errata and follow-up letters in the same section of the index, this only helps if the reader used the index to get to the material (and understood how to interpret those index entries). If the reader arrived at the original article by browsing or from a citation in another source, there would be no obvious connection to the corrections.

As for books, authors could and can write whatever they please, subject only to whatever editorial oversight the publisher exerts. The reader can look for book reviews that might criticize the information quality and compare it to other similar books, but, again, the reader needs to know how to find book reviews.

On the Web, the online publication format allows for much easier use of comments and corrections, and, indeed, this aspect is one of the great advances that Web publishing has to offer. The ease of publishing on the Net is such that if someone posts something obviously erroneous, someone else can easily post a rebuttal, refutation, or correction. Online periodicals can be sure to link corrections and letters to the original article. They can even remove or change previously published articles.

Elsewhere on the Internet, comments and links to related information are common. Discussions in Usenet news, Web forums, and mailing lists help give context, and reviews on commercial sites like Amazon and Epinions provide new information content. Weblogs offer easy content posting with the ability for others to add comments. Blogs also allow the original author to change their content. This cycle of comments, corrections, and changes is part of the changing information cycle on the Net.

WEBLOGS AND WIKIS

Prominent on many Weblogs is the opportunity for readers to add their own comments. Added to the nature of many blogs to link to other related postings, this creates a virtual community that (sometimes) provides a larger picture of an issue than any one single posting.

Consider also the Wikipedia, a collaborative encyclopedia writing project that uses wiki software to let anyone add and correct information. Active since 2001, it now has over 290,000 articles, many of which not only rank well in search engine results but also contain some quality writing and a good source for many kinds of information. It also incorporates comments under a "discussion" tab. For example, the article on Brazil has comments about the copyright status of the map used in the article and the validity of a description of one of the languages [http://en.wikipedia.org/wiki/Talk:Brazil].

Slashdot, a site for news and discussion among the technologically inclined, is a very active site with comments being a major component. The posting about the launch of the Public Library of Science [http://slashdot.org/science/03/10/13/119223.shtml] has over 100 comments from the mundane to the insightful.

EXPANSION OF CONTENT

The ability to comment and correct information can be useful in a variety of settings. Consider the typical computer software documentation. Whether in print or online, few are well written, and almost none are comprehensive. The better documentation is well-organized and goes into some depth on the program capabilities and features.

The difficulty is often that such documentation cannot include all possible errors or anticipate all questions. So why not make it a bit more interactive? The MySQL online manual with annotations does just that [http://dev.mysql.com/doc/mysql/en/]. The manual has a user comment box available at the end of each section. Previous user comments about the section are displayed along with an option to add new ones. Some comments try to clarify language. Others give examples, while a few mention situations where the program will work a bit differently than described in the documentation.

RETRAINING FOR COMMUNITY READING

Not all Internet content is published in this communal environment, nor is it necessary for many types of information. Yet for those of us used to the more bounded research process using indexes, periodical articles, and books, it is worth considering the differences with the information cycle on the Net.

When under the pressure of the clock, or the urgent user, it is easy to skim over comments, to only look at the first few results, to take the first answer presented online. Instead, I find that I am working on retraining myself to dig more deeply on the Web, to look more broadly at the range of answers, and to search for the combination of resources that gives a more knowledgeable answer. Much of that retraining involves looking at comments critically, to track links in both directions, to seek out divergent views, and to evaluate much of the content based on the Internet's information cycle rather than the print information cycle.

Comments? Email the editor at marydee@infotoday.com.