The web has greatly improved the accessibility of scientific information, however the role of the web in formal scientific publishing has been debated. Some argue that the lack of persistence of web resources means that they should not be cited in scientific research. We analyze references to web resources in computer science publications, finding that the number of web references has increased dramatically in the last few years, and that many of these references are now invalid. We also find that most invalid web references can be relocated easily. We argue that, while formal references to published articles should always be used when possible, web references help to improve communication and progress in science. However, citation practices need to be improved to minimize future loss. We provide recommended practices for citing web resources, and discuss methods for relocating invalid references.

Abstract. We explore the availability and persistence of URLs cited in articles published in D-Lib Magazine. We extracted 4387 unique URLs referenced in 453 articles published from July 1995 to August 2004. The availability was checked three times a week for 25 weeks from September 2004 to February 2005. We found that approximately 28 % of those URLs failed to resolve initially, and 30% failed to resolve at the last check. A majority of the unresolved URLs were due to 404 (page not found) and 500 (internal server error) errors. The content pointed to by the URLs was relatively stable; only 16% of the content registered more than a 1 KB change during the testing period. We explore possible factors which may cause a URL to fail by examining its age, path depth, top-level domain and file extension. Based on the data collected, we found the half-life of a URL referenced in a D-Lib Magazine article is approximately 10 years. We also found that URLs were more likely to be unavailable if they pointed to resources in the.net,.edu or country-specific top-level domain, used non-standard ports (i.e., not port 80), or pointed to resources with uncommon or deprecated extensions (e.g.,.shtml,.ps,.txt). 1

We analyze the persistence of information on the web, looking at the percentage of invalid URLs contained in academic articles within the CiteSeer database. The number of URLs contained in research papers has increased from an average of 0.04 in 1992 to 1.8 in 1999. We analyzed the validity of URLs based on the year of publication of articles. The percentage of invalid URLs in articles varied from 23% for articles published in 1999, to a maximum of 54% for 1993. We found that it was possible to find the new location of invalid URLs about 51% of the time, by either guessing (14%) or with the use of search engines (37%). For an additional 26% of URLs we were able to locate highly related information. 3% of URLs could not be easily located but were accompanied by a formal citation, while 14% of invalid links could not be easily located and were not accompanied by a formal citation (the remaining 6% were never valid). We also analyze the type of URLs cited in research articles, the correla...