Subjects

Jessica R. Buchanan

Dr. Kazmer

LIS5703 – Information Organization

December 10, 2011

Subjects

Introduction

When the goal of creating a comprehensive database is that of user findability, it is vital to take into account user expectations of information retrieval. With the increased use of online search engines, the average user anticipates certain results based upon their own understandings (Rolla). These expectations are derived from an ever growing use of folksonomies created through social tagging. In this paper, I will examine the folksonomy created by my classmates and myself, the problems that our folksonomy faced in contrast to controlled vocabulary in subject headings and how other subject access systems compare to our folksonomy in use of controlled vocabulary versus social tagging.

RefWorks Folksonomy

            Throughout the semester my classmates and I have contributed bibliographic information for resources we believed to be useful in writing our two class papers. With these contributions, we were asked to submit descriptors, or tags, to identify what our resources covered. With these descriptors, we created our own folksonomy. When this folksonomy is examined closely one can see the common issues found in social tagging.

Subjectivity in Social Tagging

Without the guidelines of a controlled vocabulary, folksonomies can “suffer from a certain degree of messiness and inconsistency” (Thomas, Caudle, and Schmitz). When one conducts a search it is based upon their understanding, and thus it stands to reason, that our descriptors would be reflective of how we interpret the articles that we contributed (Taylor and Joudrey 333) (Bates and Rowley). Folksonomies are influenced greatly by the person creating the tag and can lead to deviations and ambiguities (Thomas, Caudle, and Schmitz). In “Folksonomies: Path to a Better Way” Costentino found that the majority of tags in folksonomies were highly subjective and based on the worldview of the tagger themselves. In our class folksonomy there is a level of subjectivity based on how we came to our descriptors. While I may read an article and feel that the most important aspects are its coverage of MARC, Dublin Core, and Cataloging, a classmate may read the same article and believe that its coverage of Metadata, Cataloging Systems, and Controlled Vocabulary are the best descriptors to highlight this particular article. Without the presence of controlled vocabulary we are left to our own understandings as opposed to a set of standards that would lead to a higher level of consistency (Lu, Park and Hu).

Specific vs. General Terms

            Taylor and Joudrey mention that specific versus general terms is a challenge that even controlled vocabulary faces (336). If we examine the class folksonomy we can also see how this is a challenge faced through social tagging as well. Peter J. Rolla discusses, at great length, how social tags are generally very specific or very broad in comparison to most subject headings in controlled vocabulary. He gives a few examples of a how Library of Congress Subject Headings (LCSH) differ from how he might tag the books himself. Here is one such example: “The LC catalogs Bean’s Aegean Turkey, a guide to the archeological sites of Turkey’s western coast, under the single subject, “Ionia.” For me, however, the book is about Turkey and archaeology” (Rolla). We can see through this example that there is a gap between specific and general terms used to help find a book. In our class folksonomy we can also see the general versus specificity used when creating our descriptors. One of the best examples of this is the use of the descriptor “metadata.” Our class used the single, broad, descriptor of “metadata” 27 times in our contribute assignments. In addition to these 27 uses of the single word descriptor, there were also several more specific descriptors used in conjunction with the word metadata. Here is a list of the more specific metadata descriptors: metadata creation, metadata creation templates and editors, metadata format conversion, metadata generation, metadata harvesting, metadata implementation strategies, metadata objects description schema, metadata optimization techniques, metadata quality control, metadata schemas, metadata sharing, metadata standards, metadata structure, and metadata/evaluation. Through this we can see there were some contributors who did not feel metadata was a specific enough descriptor and thus they added to that a more precise account of what they believed their article was about.

Number of Descriptors Assigned

            For the purpose of our class contribute assignments we were given a limit to how many descriptors were to be used. Three descriptors were to be very specific and then the other three were left to our discretion for a total of six descriptors. Taylor and Joudrey touch on the topic of limited descriptors and share that this limitation can be useful in helping cut down the indexing time, but does not always benefit the user due to the fact that certain descriptors might be left out for time’s sake and may hinder the user’s ability to find what they are looking for (344). You could of course, have a similar effect if there are too many descriptors. In many social tagging environments people tend to over-tag. This leads to an overwhelming number of descriptors that can again, limit the user’s ability to find an item they are looking for by providing too many options.

Subject Access Systems

            With the increase in social tagging through sites like delicious, flickr, twitter, and many others we are left with a plethora of ways that users have become familiar with information organization. As previously stated, there are many flaws in this type of information organization, but there are also a great many things that we can learn. If we look at the class folksonomy in comparison to other subject access systems, we can see the benefit of both a controlled vocabulary and the usefulness of social tagging.

Rise of the Folksonomy

            In a time of social networking we are constantly learning and exchanging new ideas. This has led to an increase in social tagging. Wichowski touches on the idea that folksonomies grew out of a need for findability “amidst a changing environment,” and that they were “developed ‘in field’ in response to [that] environmental need.” Understanding that this type of information organization caters to a generation of users who are heavily inundated in social networking provides us with a context for the creation of folksonomies. This type of social tagging leads to a quicker, more effective means of exchanging information in a social environment leading to “an improvement of the folksonomy usage” (Nocera and Ursino).

Social Tagging vs. Subject Headings

            When we compare the class folksonomy to other subject access systems we can begin to see how controlled vocabulary in subject headings and social tagging can provide us with different types of results. If we look at The Internet Movie Database (IMDB) we see a high level of authority control and how this type of control helps with the end goal of user findability. It recognizes misspellings and offers recommendations if certain terms are not found. When you look at a site like Flickr, you can still find concrete terms, like a person, place or thing, but due to social tagging you can also find ideas, symbolisms and abstract thoughts. The problem with the lack of authority control in sites like this limits the ability to point users in the right direction if they do make an error in spelling or if what they are looking for does not exist in the system. Our class folksonomy has both authority control (Assignment, Contribute #, and Contributor’s name) and social tagging (our three self-determined descriptors). This helps to provide a view of how the two can work together and weakness that they each possess.

            Gross and Taylor discuss how “the best subject searching was done by using both natural language searching and controlled vocabulary searching in parallel.” Through their research they discovered that removing subject headings completely would actually eliminate about one-third of the results that users currently receive when doing both subject and keyword searches (Gross and Taylor). Social tagging still has the problem of inconsistency and “semantic ambiguity” which causes information professionals to stay slightly skeptic of the “value of social tagging” (Lu, Park and Hu). Yet it still stands to reason that this increase in user generated information organization can help provide information professionals with a way to “engage users with information management” (Lu, Park and Hu). Bates and Rowley also argue that the “folksonomy might be able to achieve a user-oriented retrieval aboutness.”

            Rolla states that the combination of controlled vocabulary with user tags will increase user findability. He is very adamant about the fact that user tags cannot replace the valuable use of controlled vocabularies like LCSH, but that they can “point libraries in the right direction” (Rolla). His research also showed that many user tags overlap with LCSH (Rolla). This information helps to provide us with an understanding of how user tags can help increase findability within controlled vocabulary settings such as LCSH.

Conclusion

            From the examination of the class folksonomy we created this semester we can see how controlled vocabularies in subject headings and user generated tagging can have both positive and negative effects on the findability of information objects. Lu, Park and Hu mention that “social annotation” and controlled vocabulary like LCSH would benefit from co-existence. The marriage of these two concepts would provide a more user friendly search environment that would result in better findability. Steps are already being taken to provide this type of culmination through products such as LibraryThing (Thomas, Caudle, and Schmitz). If the end product of findability is kept as a top priority among information professionals we are likely to see a more user friendly environment in which we collectively learn, and help others learn as well.

Works Cited

Antonino Nocera, Domenico Ursino, “An approach to providing a user of a “social folksonomy” with recommendations of similar users and potentially interesting resources.” Knowledge-Based Systems, 24.8 (2011): 1277-1296. Electronic.

Bates, Jo, and Jennifer Rowley. "Social reproduction and exclusion in subject indexing: A comparison of public library OPACs and LibraryThing folksonomy." Journal of Documentation 67.3 (2011): 431-48. Electronic.

Cosentino, Sharon L. "Folksonomies: Path to a Better Way?" Public Libraries 47.2 (2008): 42-7. Electronic.

Gross, Tina, and Arlene G. Taylor. "What Have We Got to Lose? The Effect of Controlled Vocabulary on Keyword Searching Results." College & Research Libraries 66.3 (2005): 212-30. Electronic.

Lu, Caimei, Jung-ran Park, and Xiaohua Hu. "User tags versus expert-assigned subject terms: A comparison of LibraryThing tags and Library of Congress Subject Headings." Journal of Information Science 36.6 (2010): 763-79. Electronic.

Rolla, Peter J. "User Tags versus Subject Headings: Can User-Supplied Data Improve Subject Access to Library Collections?" Library Resources & Technical Services 53.3 (2009): 174-84. Electronic.

Taylor, Arlene G. and Daniel N. Joudrey. The Organization of Information. 3^rd ed. Connecticut: Libraries Unlimited, 2009. Print.

Thomas, Marliese, Dana M. Caudle, and Cecilia Schmitz. "Trashy Tags: Problematic Tags in LibraryThing." New Library World 111.5-6 (2010): 223-35. Computer and Information Systems Abstracts. Electronic.

Wichowski, Alexis. "Survival of the fittest tag: Folksonomies, findability, and the evolution of information organization." First Monday (Online) 4 May 2009. Electronic.