James Ho's Research Brief in this issue constructively raises a number of issues of current practical and academic interest to the editors of the Journal of Computer-Mediated Communication. The challenges we faced in migrating the JCMC files to a new server at the beginning of this year have made us acutely aware of the problems of preserving an electronic journal. Broken links, both within and directed outside the JCMC site, are one of the problems we face.
Ho offers three recommendations regarding the maintenance of hyperlinks in electronic journals. The first suggests that "precision standards" be adopted for hyperlinks to be meaningful, i.e., that they direct a user to a specific page where the requisite information is found. In other words, authors—and ultimately, journal editors—have a minimum responsibility for the meaningfulness of links, as they do for traditional references. This seems reasonable. The second and third recommendations propose editorial responsibilities that continue after publication, through a mechanism for reporting and decommissioning broken links. This suggestion seems problematic in several ways.
First, it effectively places archival concerns in the hands of journal editors, where they arguably do not belong. The institutions that host electronic academic materials—schools and departments in universities, private-sector corporations, etc.—typically do not have long-term commitments that include archiving. When editors change, the institution's involvement may end, domain names may change, and hitherto available materials may disappear. Let us imagine, however, that the archiving problem has been solved, and that JCMC and other online journals are preserved for posterity. If dead links must be deactivated, electronic journals would require continuous editorial maintenance in whatever archive they are stored. This means that articles would continue to be "dynamic" throughout their archival lifespan, changing after publication, and complicating the challenge of archiving. Moreover, if other sites archive the content of JCMC, as recommended for example by the LOCKSS (Lots of Copies Keep Stuff Safe) initiative (http://lockss.stanford.edu/), which version of an article would be definitive? If links go dead, do mirror archives also need to be updated? Where does the updating happen, and how does one ensure that the mirror and original sites have common content? In short, the proposal to decommission broken hyperlinks entails complications and costs. Would the benefit to readers be worth the price?
Second, broken links are not unique to academic electronic publishing; they are a fact of life on the modern Web. While 404 error messages, redirects and other symptoms of broken links are annoying, keeping the link active makes it easier to copy the link and paste it into a search engine to get further information, e.g., from the Internet Archive (http://www.archive.org/), or from a search engine's link sets. The Internet Archive is particularly helpful from the perspective of Ho's "connection" purpose, and shows the importance of good archiving. For example, as of this writing, the main page of the "Arc Essays" (http://www.arc.org/essays.html), referenced by a hyperlink in the references section of the first article published in JCMC 1(1) (Acker, 1995), returns an HTTP 404 error, and should be de-activated according to Ho. Searching in the Internet Archive, however, turns up a viable link on February 8, 1997, which can be browsed to reveal the referenced essays. Hence, all of the original content of the reference remains available through the Internet Archive. Having the active link is a convenience, in this circumstance, as it facilitates searching the archive. One can also search Google's indexes using the search text "link:www.arc.org/essays.html" to discover that Google knows of two sites still referring to this page, which tells us something more about the status of the broken links. Both of these kinds of searches could readily be implemented as browser toolbar features in Mozilla or Internet Explorer, so dead-link checking could be made more much transparent than it presently is.
In light of these observations, it seems to us at JCMC that link-checking should not be added to the already long list of responsibilities of the editors of an electronic journal. As a task that has no fixed end-point, it is outside the traditional responsibilities of editors, whose primary responsibility in citation is to ensure the credibility and credit purposes. Moreover, users should not overlook resources like the Internet Archive, and Web browsers could readily be configured to support link-checking. Ho's Research Brief draws needed attention to the problem of broken links, but we feel that workable solutions ultimately lie elsewhere, in digital archive initiatives and the development of better online search tools. JCMC encourages and supports efforts to advance these frontiers, not only for the sake of online scholarly journals, but for the future of digital documents in general.
Acker, S. R. (1995). Space, collaboration, and the credible city: Academic work in the virtual university. Journal of Computer-Mediated Communication, 1 (1). Retrieved April 5, 2005 from http://jcmc.indiana.edu/vol1/issue1/acker/ACKTEXT.HTM and http://jcmc.indiana.edu/vol1/issue1/acker/ACKBIB.HTM
Arc essays:
http://www.arc.org/essays.html
http://web.archive.org/web/*/http://www.arc.org/essays.html
http://web.archive.org/web/19970208235706/http://www.arc.org/essays.html