Creating metadata that work for digital libraries and Google

Abstract

For many years metadata has been recognised as a significant component of the digital information environment. Substantial work has gone into creating complex metadata schemes for describing digital content. Yet increasingly Web search engines, and Google in particular, are the primary means of discovering and selecting digital resources, although they make little use of metadata. This article considers how digital libraries can gain more value from their metadata by adapting it for Google users, while still following well-established principles and standards for cataloguing and digital preservation.This article introduces the concepts of functional and variable metadata, and explains why they may be of value to users and managers of digital libraries that rely on Web searching as a significant means of resource discovery.Functional means something that works, so "functional metadata" is used here to mean metadata that fulfils its primary function of assisting information retrieval. Not all metadata does this in a Web-based world.Variable means something that may change, so "variable metadata" is used to refer to metadata that may vary according to context. This is not the same as "dynamic metadata", which has been used to describe educational metadata that can influence the behaviour of multimedia learning objects (El Saddik, 2000).In order to consider why functionality and variability might be useful qualities for metadata, it is necessary to acknowledge the dominance of the Web and of Google as means of access and resource discovery for digital libraries. The current pre-eminence of Google extends well beyond the Web: a recent survey, drawing on users from 85 countries, rated Google as the world's number one brand name, above Apple, Mini, Coca-Cola, Samsung, Ikea and Nokia (Brandchannel.com, 2004). One might think information professionals would be delighted that an information retrieval tool had become the world's leading brand. However, some librarians have been known to denigrate Google because it "doesn't work". Given that it can search millions of documents for thousands of users simultaneously, and deliver useful results within seconds, this is clearly a specialist interpretation of "doesn't work". Yet one can understand the sentiment. Many of the things that librarians take for granted simply are not possible with Google. Trying to find articles written by Tony Blair, as opposed to those written about him, is difficult. A library catalogue system would make this easy, as "Blair, Tony" would be entered in the author field. But in a library system items are not normally catalogued at the article level, so the search might produce zero hits even though the retrieval system worked perfectly.This basic problem illustrates the need for functional metadata (and the value of article-level retrieval). In the past cataloguers have been able to concentrate on capturing the metadata of an

Collaborative tagging as a knowledge organisation and resource discovery tool

“…Metadata can enhance the process of resource discovery by disclosing sufficient information about a resource to enable users or intelligent agents to discriminate between what is relevant and what is irrelevant to a specific information need. Metadata also facilitates approaches to searching or browsing that are simply unfeasible using existing post-coordinate systems (Dawson, 2004). For example, it enables the location of resources on the subject of 'Adam Smith', rather than those written by 'Adam Smith'.…”

A Study of HTML Title Tag Creation Behavior of Academic Web Sites

“…The title of web pages may be changed by webmasters at any time, while traditional library materials, such as books, may have a subtitle or alternative title as well as a main title, but these never change after publication. 20 A uniform title for each web page is necessary as an access point. However, the procedure for determining the uniform title for a web page is not always simple and the title of web pages may not always reflect the true content.…”

“…The title tag is the only effective metadata element in HTML, and the primary means of making metadata functional via the Web. 4 Users have to browse numerous web pages in their search results to identify relevant items. It therefore follows that the title tag is vital for web pages.…”

Optimising Publications for Google Users

“…To pursue the above example of the Journal of Internet Cataloging, the main reason that Google search results were so relevant was because the publishers had helpfully ensured that the journal title appeared in the title tag (so they had done part of their job well). Some further guidance on use of the title tag to enhance Google searching is given by Dawson (2004).…”