1 Dublin Core is Still Dead / By Jeffrey Beall Note: This is a post print of an article that was published here: Jeffrey Beall, (2014),"Dublin Core is still dead", Library Hi Tech News Vol. 31, Issue 9, p. 11 13. Permanent link to this document: http://dx.doi.org/10.1108/LHTN 07 2014 0058 Introduction In 2004, Library Hi Tech News published my article "Dublin Core: An Obituary" (Beall, 2004). Classified as a viewpoint article, the piece was a reaction to the efforts of some in the academic library and library consulting communities to kill off the MARC standard and replace it with the Dublin Core metadata standard. Despite the MARC format's opponents' intensive and enduring efforts, Dublin Core is now essentially obsolete, and libraries still extensively use the MARC standard in their databases and discovery tools, creating new MARC records all the time. In this article I reflect on my original 2004 article, its impact , and the long, slow, and agonizing death of Dublin Core (DC) over the past decade. The Dublin Core metadata standard appeared in 1995, a time when it was trendy and cool to name new standards after the location where they were conceived. In this case, the location was Dublin, Ohio, a ritzy suburb of Columbus and the home of OCLC. In the inchoate World Wide Web of 1995, it was often difficult to find desired information, for Google and its advanced search algorithms had not yet appeared. The best search engine at the time was Alta Vista, and even it was unreliable. Metadata experts wanted to help improve searching on the Web, but they were ashamed of the MARC format and its accompanying content standards and badmouthed it Â—
2 forestalling any possibility of its application to web standards Â— to those who were developing and advancing HTML standards. They wanted a new and au courant standard, one simple enough that any website developer could use.  In the academic library world of the mid 1990s, Dublin Core was a very big deal. After its development in early 1995, Dublin Core was touted as the metadata format of the Web Â— indeed of the future. Its promoters promised that it would revolutionize the discovery of networked resources (as they were called then), and the standard would be dumbed down (though they didn't say it that way) so that any website creator could easily create the metadata search engines would use to index the websites, improving search results' precision and recall. Academic libraries made grand preparations for the advent of the Dublin Core standard, rearranging their workflows and staffing in preparation, but it never really arrived. They also spent untold funds on training and workshops, enriching the purses of the consultants who ran the workshops (and who also rigorously promoted the standard). The Dublin Core Metadata Initiative, the organization behind the standard, aimed to be a truly international organization and demonstrated this by holding its annual meetings in far flung places like China, further stretching the budgets of libraries sponsoring the travel costs of their staff members who attended. Why I Wrote the 2004 Article I wrote the article chiefly because most all the information about Dublin Core and the arguments supporting it were one sided and biased. I sought to provide a more balanced and honest view of the utility of Dublin Core and its potential for success as a metadata standard. But because there was a strong measure of political correctness supporting the standard, most were afraid to speak openly about its weaknesses
3 for fear of being stigmatized by the DC supporters. Now ten years later the standard is all but abandoned, an indication that more critical thinking about Dublin Core was justified and needed. It is, unfortunately, a part of library culture to quickly accept technological novelty without sufficient analysis or forethought. For example, writing in 2003 about the shift from pre coordinated subject searching to keyword searching, Library of Congress reference librarian Thomas Mann (2003) wrote: I strongly suspect one reason lies in our profession's characteristic inferiority complex. If anything smacks of being a carryover from the age of manual card catalogs, we're horrified at the prospect of being considered old fashioned (p. 54). Similarly, when the Resource Description and Access (RDA) standard was being introduced, Michael Gorman (2007) said: Articles stating that RDA will be dead on arrival have already appeared, not because it is a mess and a giant leap backwards for cataloging, but because the neophiliacs think it is not a radical enough break with the past. (p. 65). These attitudes among librarians were exploited by the promoters of Dublin Core to help pressure libraries to implement the standard even though it was clearly not sufficient for the type of high quality information retrieval that is needed in academic and research libraries. One of the people who most vociferously trashed the MARC standard was Roy Tennant. He had a regular column in the magazine Library Journal and unprofessionally used it to mock
4 the MARC standard. For example, he wrote columns with titles such as "MARC exit strategies" (Tennant, 2002) and "MARC must die" (Tennant, 2002b). I think he wrote the pieces not so much to share thoughtful comments on MARC but to impress his friends and to draw attention to himself as a crusader against the MARC standard. Tennant later began working for OCLC, a company that earns much of its revenue re selling MARC records created and contributed collectively by librarians from all over the world. Upon accepting this position, he immediately stopped bashing the MARC standard, which was now helping pay his salary. This prompted one academic librarian to write: Roy Tennant declared that "Marc must die" in 2002, although he has since retreated from that viewpoint and adopted one that incorporates MARC into the broader context of metadata standards" (Wakimoto, 2009, p. 410). The title of my piece, "Dublin Core: An obituary," was meant to counterbalance the death motif first expressed in the titles of Tennant's columns attacking MARC in Library Journal The Impact of the article The article had an immediate impact. Soon after its publication, I received an invitation to speak at the Second Canadian Metadata Forum held at Library and Archives Canada in 2005. The government of Canada, had invested heavily on the promise of Dublin Core and mandated that the standard be applied in all government websites. However, the implementation of the standard was progressing poorly. In fact, it was a disaster, requiring much effort with little actual return on the investment. Many Canadian government websites contained the required Dublin Core fields in the website headers, but they were not populated with any
5 meaningful metadata; they were empty. My article resonated with the Canadian government officials frustrated with DC's weaknesses. I turned the text of my Canadian Metadata Forum talk into a scholarly article; it was published with the title, "The Death of Metadata" (Beall, 2006), continuing the death motif. This was actually the title of my talk at the forum, a title chosen by the organizers. The article records some of the anguish expressed by Dublin Core promoters upon the publication of my 2004 "obituary" article, which coincidentally was published online just as the International Conference on Dublin Core and Metadata Applications (DC 2004) was convening in far flung Shanghai, China. Technology Change and Consultants I think the best technological change happens when it occurs naturally and not when it's forced by a social movement or by consultants, as was the case with Dublin Core. The Dublin Core Metadata Initiative had some of the characteristics of a social movement, not unlike the contemporary open access movement. I wrote about this recently: In recent decades in my field, library science, we have lived through a giant technological and social change that greatly affected the operations of almost every library in the world: the transition from card catalogs to online catalogs. In this case, the change didn't have a social movement behind it; it was organic. There was no one shouting "Card catalogs must go," or "Online catalogs now!" The change occurred naturally, with everyone seeing the advantages of online catalogs, implementing them as soon as they could. Card catalogs were gradually phased out, the newer technology and practices replacing the old (Beall, 2014, p. 5).
6 Not only do social movements aim to force technological change; consultants do also. I think that many of the people pushing for the broad implementation of Dublin Core were consultants who would directly benefit from such an implementation. They were the ones providing the training, coding and implementation of Dublin Core for many libraries. Many of them came from the East Bay area of California. The University of California System used to operate a union catalog called Melvyl. Like a number of consortial online public access catalogs (OPACs) in the early days of online catalogs, Melvyl was developed in house by programmers and database designers who were employees of the university. In the early 2000s, the University of California finally decided to terminate the in house development of the OPAC, opting instead to use an off the shelf product, that is, one provided by a vendor. Some of Melvyl's developers were laid off throughout the transition, and it almost appears as if they sought to take out revenge on MARC because of the mothballing of the original Melvyl catalog, as if the standard was a nemesis. In fact, some of those made redundant by the shelving of Melvyl are still wandering around Western countries speaking on the future of libraries, a never ending, Melvyl created diaspora. Others work as consultants. Reflecting on the history of DC, I question the nature of consulting work mixed with standards advocacy in library technology. I think there needs to be a code of ethics for library and information technology consultants, many of whom effectively use social media to create artificial needs for the services they provide, much like soap companies using advertising to convince consumers they need a particular type of new soap. A group of consultants can effectively sway library public opinion through social media and conference presentations, creating new business for
7 their consultancies. This is a conflict of interest, and I think we need a code of ethics for library and information technology consulting. Conclusion In the end, here's what really killed Dublin Core: Schema.org, the metadata standard launched by the search engine industry in mid 2011. The launch essentially made Dublin Core obsolete, at least in the context of the web environment. It was a rejection of the Dublin Core metadata standard by the industry the standard's proponents most wanted to impress, a fatal blow if there ever was one. Therefore, Dublin Core is still dead. Notes . According to Google Scholar as of July 11, 2014, the article has been cited 16 times. . MARCXML can be included within HTML just as easily as Dublin Core. References Beall, J. (2004), "Dublin Core: An obituary", Library Hi Tech News, Vol. 21 No. 8, pp. 40 41. Beall, J. (2006), "The death of metadata", The Serials Librarian Vol. 51 No. 2, pp. 55 74. Beall, J. (2014), "Unintended consequences: The rise of predatory publishers and the future of scholarly publishing", Editorial Office News Vol. 7 No. 1, pp. 4 6.
8 Gorman, M. (2007), "RDA: Imminent debacle", American Libraries, Vol. 38 No. 11, pp. 64 65. Mann, T. (2003), "Why LC subject headings are more important than ever", American Libraries Vol. 34 No. 9, pp. 52 54. Tennant, R. (2002), "MARC exit strategies", Library Journal Vol. 127 No. 19, p. 27. Tennant, R. (2002b), "MARC must die", Library Journal Vol. 127 No. 17, p. 26. Wakimoto, J.C. (2009) "Scope of the library catalog in times of transition", Cataloging & Classification Quarterly Vol. 47 No. 5, pp. 409 426.