Abstract

In recent years the authors of scholarly publications have relied on e-resources. But e-resources have raised the question of permanency on the web. In this context, this article investigates the availability, persistence of Uniform Resource Locator (URL) citations cited in two Library and Information Science (LIS) journal articles published by Emerald Publishers during 2008 and 2012. In total, 2477 URLs cited in 406 research articles published in two LIS journals spanning a period of five years (2008–2012) were extracted. The study found that 23.81 per cent (2,477 out of 10,400 references) of URLs were cited in these journal articles. 49.53 per cent of URL citations were not accessible and the remaining 51.47 per cent of URL citations were still accessible. The study used W3C link checker to identify HTTP errors associated with missing URLs. HTTP 500 error message—‘page not found’ was the overwhelming message that represented 39.18 per cent of all HTTP error messages. This study attempts to focus on URLs link rot and its implications for electronic publishing. 

Publishing segment has been greatly influenced by Information and Communication Technology. E-publishing is a remarkable change in the field of publishing. Scholarly world has witnessed a shift from conventional publishing to electronic publishing. This is because e-publishing has reduced the cost and time that are required for conventional publishing. Ever since the first e-book was published in 1985, electronic publishing has seen steady growth (Moorthy and Karisiddappa 1996). Today, millions of electronic resources are available and being published over the World Wide Web (WWW). E-publishing has also influenced the convenience of researcher in information search and retrieval. Despite its convenience, recent studies have documented the problem of loss of web resources or URL link rot. It is a serious issue for web masters and academic community (Dimitrova and Bugeja 2007). Hence, this article attempts to identify the permanence and URL link rot by studying the URLs cited in two LIS e-journals published by Emerald Publishers. 

• To know the extent of the use of web resources in LIS scholarly communication.


• To know the top-level domains and file formats associated with URL citations.


• To check the accessibility and decay of URL citations cited in LIS scholarly communication.


• To know the HTTP error messages associated with missing URL citations.

The data for the present study was drawn from a selective sample of LIS scholarly journals published by Emerald Publishers. The following two journals were selected for the current study:

• Program: Electronic Library and Information Systems


• The Electronic Library

The time frame for the analysis was the publication years 2008–2012. All research articles published during the 5-year period were downloaded and saved. In totality, 10,400 references were selected from 406 articles published in the selected journals. Only those references that appeared as a list at the end of the article under the references section were considered. Editorial notes or book reviews, expanded bibliographies, end notes, foot notes, e-mail links, and annotations were not considered and were not tested or counted in our dataset.

The study aims to check the availability of URL citations and their persistence. Duplicate URLs were recognized and removed from the list. W3C link checker (http://validator.w3.org/checklink) was used to check the URLs’ existence (Sellitto 2005). After the initial check, the URL citations were grouped as ‘active’ and ‘missing’ URL citations. The missing URLs were entered in the Wayback machine and the ‘Take Me Back’ button was clicked to retrieve the missing URLs from the Wayback machine.

It is evident from Table 1 that the percentage of URL citations is not consistent during 2008–2012. The percentage of URL citations varied from the highest of 29.78 per cent in 2009 to a lowest of 18.06 per cent in 2012. It is also an observable fact that the percentage of URLs per article is above 5 per cent.

Table 2 illustrates the summary of active and missing URL citations. The percentage of missing URL citations varied from a highest of 69.15 per cent in 2009 to a low of 33.61 per cent in 2012. A total of 49.53 per cent of URL citations were not accessible during the initial check of URLs in W3C link checker. It is evident from the table that as the age of the URLs increases their permanency is questionable. The decreasing trend of missing URLs from 2009 to 2010 could be witnessed in Table 2 and Figure 1.

Around 1,202 missing URL citations were entered in the Wayback machine and the ‘Take me back’ button was clicked. About 40.60 per cent of missing URLs are recovered back from Wayback machine. The search in the Wayback machine yielded a total of 488 URLs that have been retrieved successfully (Table 3).

The ‘HTTP 500 error message’ was the overwhelming message encountered and represented 39.18 per cent of all HTTP error messages, followed by HTTP 404 (36.52 per cent) and HTTP 403 (12.56 per cent). It is interesting to note that 43.74 per cent of URLs were recovered from HTTP 500 error message and 43.05 per cent of missing URL citations were recovered from HTTP 403 error message. 36.22 per cent of missing URLs with HTTP 404 error message were recovered from the Wayback machine.

The top-level domain having the greatest number of missing URLs was the .gov domain (67.33 per cent) followed by .edu (57.09 per cent). It is notable that a low level of URL link rot is associated with the .net (29.91 per cent).

Table 4 illustrates the distribution of different domains associated with missing and retrieved URLs. URLs with .com/.co (45.24 per cent) and geo domains (44.31 per cent) are the highest retrieved missing URLs followed by URLs with .org (42.18 per cent) and .net (37.14 per cent) domains. It is interesting to note that the URLs with academic (30.34 per cent) and educational (34.32 per cent) domains are the least recovered URLs.

The data as illustrated in Table 5 indicates that the greatest numbers of cited URLs in LIS scholarly communication web resources with HTML/HTM files (65.48 per cent). Out of 2,477 URLs, 1,622 cases are HTML files, followed by 689 (27.82 per cent) that are PDF files and 112 that are PHP files (4.52 per cent).

Table 5 also depicts the distribution of recovered URLs having different file types. URLs with PDF file types (41.21 per cent) and HTML file types (40.93 per cent) recorded highest recovery whereas URLs with .jsp file types and .doc recorded 15.38 per cent and 23.08 per cent respectively, which is a very low recovery rate. Missing URLs with PDF files format showed highest recovery than of other file types.

The data presented in Figure 2 illustrates the missing and recovered URLs with different path depths. It is clear from the analysis that higher the path depth of the URLs, less is the recovery whereas, lower the path depth of URLs, higher the recovery. Path depth-1 records 30.43 per cent of recovery whereas URLs with path depth of 8 recorded very less recovery, i.e., 15.78 per cent. URL with path depth of 4 and less shows highest recovery, whereas URLs with the path depth of 5, 7 and above recorded less recovery.

The distribution of URLs by character length is given in Table 6. It illustrates the percentage of missing URLs and recovered URLs from various URLs having number of character length. The highest percentage of missing URLs (61.30 per cent) was found in the URLs having 51–60 characters followed by >70 character length (55.48 per cent).

It is clear from the analysis of Table 6 that the highest character length of URLs lowers the recovery rate. The data indicated in Table 6 that the Wayback machine could recover only 24.41 per cent of missing URLs with the character length of >70 followed by 37.50 per cent of URLs with 61–70 character length which recorded lowest recovery rate; whereas, the recovery is highest among the URLs with character length of 31–40 (53.33 per cent) and 21–30 (49.31 per cent). 

E-publishing has gained momentum in the scholarly world. The low cost to produce information electronically and the speed of publication have made it an inevitable source in the scholarly environment. E-publishing has given opportunity to post any kind of information on the Internet. But, it has also raised the question of unavailability of web resources in the longer duration. Several studies have shown that URLs of web pages would disappear as their age increases (Wren et al.; Aronsky et al. 2007; Dimitrova and Bugeja 2007; Ducut et al. 2008; Goh and Ng 2007; Sampath and Vinay 2013). Since majority of web sources vanished from the original web location due to various reasons, the users of the web find it very difficult to see the needed web sources. There are many reasons why online sources disappear. Some websites simply cease to exist because their website domain names expire. Other web pages are removed from the web by the website creator. Some websites are redesigned with new file structures (Dimitrova and Bugeja 2007). In the present study we found that 48.53 per cent of URL citations were not accessible and the remaining 51.47 per cent of URL citations were still accessible. The HTTP 500 error message—was the overwhelming message encountered and represented 36.52 per cent of all HTTP messages. The study also checked the rate of URL link rot associated with the URLs with different domains. URLs with .gov domain were found to be highest missing which represents 67.33 per cent of the total followed by .org domain (57.09 per cent). It is a considerable observation of the study that the electronic publication from governmental and organizational domains are most liable to disappear.

Many studies (Dimitrova and Bugeja 2007; Wren et al. 2006) have already recommended various strategies for the publishers, editors, and authors in order to increase the rate of availability of URLs. Wren et al. opined that loss of URLs would be continued until better preservation policies are adopted. Authors are in a vulnerable position to maintain the web pages who post content on the Internet. They are often ill equipped to maintain it (Sampath and Vinay 2013). But authors may give parallel citations to e-resources that are referred in their scholarly work. In the long term, all research organizations must come to the realization that, due to the dynamic nature of the Internet, certain minimum standards must be maintained in posting scholarly information there (Altman and King 2007). It is recommended to publishers of electronic web pages that they have to maintain lower path depth in the URL address, because several studies identified that lower the path depth lesser the missing citations (Goh and Ng 2007; Sampath and Vinay 2013). Apart from this, the Digital Object Identifier (DOI) system could be used with the URLs of electronic resources. It is considered as PURL—Persistence Uniform Resources Locator. It would be solution to disappearing act of web pages (DOI.org 2012).

Altman M and King G. 2007. A proposed standard for the scholarly citation of quantitative data. D-LibMagazine. Available at <http://www.dlib.org/dlib/march07/altman/03altman.html> (last accessed on June 25, 2013).

Aronsky D, Madani S, Carnevale R J, Duda S, and Feyder F T. 2007.  The prevalence and inaccessibility of Internet references in the biomedical literature at the time of publication. Journal of the American Medical Informatics Association 14(2): 232–234.

Dimitrova D V and Bugeja M. 2007.  The half-life of Internet references cited in communication journals. New Media & Society 9(9): 811–826.

DOI.org. 2012. DOI System and Persistent URLs (PURLs). Available at <http://www.doi.org/factsheets/DOI_PURL.html> (last accessed on  July 1, 2013).

Ducut E, Liu F and Fontelo P. 2008. An update on uniform resource locators (URL) decay in MEDLINE abstract and measures for its mitigation. BMC Medical Informatics and Decision Making 8: 23. Available at <http://www.biomedcentral.com/1472-6947/8/23> (last accessed on July 5, 2013).

Goh D H L and Ng P K. 2007. Link decay in leading information science journals. Journal of the American Society for Information Science and Technology 58(1): 15–24.

Moorthy L R and Karisiddappa C R. 1996. Electronic publishing: Impact and implications on library and information centres. pp. 15–34. In Digital Libraries: Dynamic storehouse of digitized information, edited by N M Malwad, T B Rajashekar Rao, I K Ravichandra and N V Satyanarayana, New Delhi: New Age International.

Sampath Kumar D and Vinay Kumar B T. 2013. HTTP 404-page (not) found: Recovery of decayed URL citations. Journal of Informetrics 7: 145–157. Available at <http://dx.doi.org/10.1016/j.joi.2012.09.007> (last accessed on June 29, 2013).

Wren J D, Johnson K R, Crockett D M, Heilig L F, Schilling L M and Dellavalle R P. 2006. Uniform resource locator decay in dermatology journals: Author attitudes and preservation practices. Archives of Dermatology 142: 1147–1152.

Sellitto C. 2005. The impact of impermanent web located citations: A study of 123 scholarly conference publications. Journal of the American Society for Information Science and Technology 56(7): 695–703.