You are on page 1of 13

!"##$#%&'()*+&,-.

/0/1234&-05&6/784/0/1234&

&

&

&9&

Taxonomies and Folksonomies in Search and Retrieval Ashley Swan, Katie Greving, Greta Kliewer Emporia State University

LI804XR: Organization of Information Prof. Kalyani Ankem December 1, 2012

!"##$#%&'()*+&,-./0/1234&-05&6/784/0/1234&

&

&

&:&

Introduction Today, more information is available to users than has ever been available before. The advent of digitization and digital creation of information has led to the explosion of online resources available to those with Internet access, and it is one of the core responsibilities of libraries and information institutions to provide patrons with access to this information, so their questions can be answered. However, while more information is available than ever before, its sheer volume can make navigation overwhelming, and does the user no good if the proper resources cant be retrieved. The need to organize this digital information has led information professionals to find ways to update and strengthen traditional retrieval methods. Information professionals throughout history have created methods for not only organizing and archiving information, but for alerting patrons to its content and making it available for easy discovery. This paper discusses two primary theoretic methods by which this is accomplished: taxonomies and folksonomies. The paper details the history of both methods, their applications, and where they fit into the future of information retrieval. It also discusses a few perhaps unintended outcomes of using tags within a folksonomy. It is clear from the literature that both taxonomies and folksonomies present powerful opportunities for arranging and retrieving information to fulfill patrons needs. History and Definition of Folksonomies and Taxonomies Taxonomies The origins of some form of taxonomy can be traced as far back as the ancient Greeks and their library at Alexandria. Egyptian wall paintings dating back to 1500 BC also illustrate and classify different varieties of medicinal plants and the diseases they treat. The Greek philosopher Aristotle was the first to classify all living things, and some of his groupings still exist in the biological classification system used today (Manktelow, p. 1-2, 2010). In more modern history, Carl Linnaeus invented a form of taxonomy to classify botanical, and then later, animal specimens. The classical theory of categories has evolved in the last sixty or so years. Previously, categories were defined as mutually exclusive groups, that is, things could belong to one and only one category, and the shared, common attributes of the group members were what defined the category (Taylor & Joudrey, 2009). Now, we tend to view categorization in a much broader context, where there are more grey areas and items can be categorized in multiple ways. Charles Cutter was one of the first to explore the use of taxonomy to improve subject access in libraries. In 1876, Cutter published his book, Rules of a Dictionary Catalog, and it served as a precursor to the ALA List of Subject Headings, published nine years later in 1885. The subject headings were derived from the classification systems within five of the major US libraries at the time, including the Boston Athenaeum, where Cutter worked (Taylor & Joudrey, p. 77, 2009). Library of Congress Classification originated when the Library of Congress burned in August 1814 after an attack by the British, and Thomas Jefferson offered his 6487-volume collection to be purchased by Congress as the beginnings of a new library (Miksa, p. 3, 1984). Jefferson had organized his library into 44 main classes; however, it became clear that a more detailed classification was needed. The Library of Congress drew heavily on Charles Cutters classification system, called Expansive Classification, which classified subjects by letter and then by number (Taylor & Joudrey, p. 78-79, 2009). The Library of Congress Classification is hierarchical; it begins with broad, top-level categories, which branch into any

!"##$#%&'()*+&,-./0/1234&-05&6/784/0/1234&

&

&

&;&

number of subordinate levels, moving from the general to the specific (Taylor & Joudrey, p. 384, 2009). It is also very enumerative, meaning that it designates a classification for any possible topic within its hierarchy, rather than branching out through faceting.

The use of controlled vocabulary taxonomy in libraries can be traced back to the creation of the Library of Congress subject headings. In the early 1900s, the Library of Congress created the beginnings of what became the Library of Congress Subject Headings. As new terminology flourished in the early 1900s, librarians asked for a list of the Library of Congresss subject headings to serve as a basis for their own. The first edition of the Subject Headings Used in the Library of Congress was published in 1914. It is now updated weekly in its online version (Taylor & Joudrey, p. 78, 2009). As information professionals found this system more and more conducive to their organization needs, they began implementing it on resources that their libraries already held: By the 1930s more and more libraries found themselves doing a retrospective conversion of subject headings. They abandoned their homegrown lists, the outdated (1911) ALA List, or Sears, and converted to the LCSH for several compelling reasons: as their library collections grew, the desire for more precise subject headings increased; it was no longer economically feasible to continually revise subject headings appearing on LC catalog cards; and, the LCSH was the only general list that made a consistent practice of keeping up to date by creating new subject headings for new topics (Stone, 2000). This was a source of controversy at times as professionals were forced to switch from taking a proprietary view of the LCSH, to a more cooperative one, serving all libraries - not just their own. Another common form of taxonomy classification that has been widely used in the United States is the Dewey Decimal Classification system. The DDC was first published by Melvil Dewey in 1876 and named ten main classes of study, ten divisions within each class, and ten sections within each division, using the numbers 000 999 as the basis for classifying its topics. In this way, Dewey Decimal Classification is also hierarchical, moving from broad classes to very specific concepts. Its primary use is different from the LCSH. The DDC uses a hierarchical alphanumeric classification to help identify and find a resource. Its index arranges topics in alphabetical order. The DDC is considered more flexible than the LCSH, but is less open to the addition of new subjects. Within this framework, however, further subdivisions make the topic increasingly more specific by adding more decimal points to the call number. This makes Dewey Decimal Classification somewhat faceted: it attempts to divide the universe of knowledge into its component parts, and then to gather those individual parts into categories or facets (Taylor & Joudrey, p. 387, 2009). Within each broad class, it seeks to include many related topics to create a complete picture of that concept. Taxonomies, used primarily in libraries, were once used more commonly on the Web in the form of online directories, such as Yahoo! Directory and Open Directory Project. However, as search engines and folksonomies have become more popular, many of these online taxonomies, such as the now-defunct Zeal Directory, have fallen by the wayside (Hammond, Hannay, Lund, & Scott, p. 3, 2005). Folksonomies Kroski defines a folksonomy as a non-hierarchical ontology that is created as a natural result of useradded metadata or tagging (2009). Cosentino uses the word folksonomy not just to describe the classification system itself, but also the phenomenon of users generating metadata by tagging pieces of digital information in their own searchable keywords (p. 42, 2008). Cosentinos definition seems significant because of the central role that users and user tags play in shaping a folksonomy.

!"##$#%&'()*+&,-./0/1234&-05&6/784/0/1234&

&

&

&<&

Information architect Thomas Vander Wal coined the term folksonomy in 2005 by blending the words folks, meaning people, with taxonomy to describe the online tagging system that employs an entirely new source of data for finding and organizing information: user participation (Taylor & Joudrey, p. 366, 2009). Folksonomies can go by many names, including folk classification, collaborative tagging, social bookmarking, and collective vocabulary. Many websites currently make use of folksonomies as a usercentered format for organizing resources. Through folksonomies, users can tag web pages through Delicious, videos through YouTube, photos through Flickr, academic papers through CiteULike, books through Goodreads, or products through Amazon, among many other sites. The concept of saving online resources for later use, a primary purpose of any folksonomy, began in with the advent of Mosaic, the first modern Internet browser, in 1993. Mosaic was developed by the NCSA, or National Center for Supercomputing Applications, at the University of Illinois. It organized and managed its links systematically in a hierarchy rather than simply listing them all on its home page (Hammond, Hannay, Lund, & Scott, p. 3, 2005). Mosaic later developed a feature called Hotlists, which mimicked the file system of folders and files to organize hyperlinks. One of the writers of Mosaic, Marc Andreesen, continued his work on Mosaic after graduating from University of Illinois, and he released its successor, Netscape Navigator, in 1994. Within the Netscape browser, the Hotlists feature was renamed Bookmarks. Microsofts web browser, Internet Explorer, included a similar link feature called Favorites. However, the popularity of bookmarks was temporarily dwarfed by the arrival of search engines like Yahoo! (launched in 1995) and Google (launched in 1997). It soon became just as easy to use a search engine repeatedly as it was to look up a bookmark (Hammond, Hannay, Lund & Scott, p. 3, 2005). The World Wide Web Consortium began work on its Annotea Project in March 2001. The consortium, also called W3C, is led by Tim Berners-Lee, who is credited with the creation of the Internet. Annotea allowed users to bookmark websites, annotate them, and create a comment thread to comment on others bookmarks (Facts about W3C, 2012). Annotea set the stage for current social bookmarking sites with folksonomies, including many of the ones in use today. Folksonomies permit actors to describe documents with subject headings called tags, without regard for conventional rules (Keshet, 2011). Tags become larger (in font size in tag clouds) as they become more widely used:

Figure 1: Tag Cloud Traditional classification systems use controlled vocabularies to create headings. In folksonomies, tags replace subject headings, which instead of being selected by a group of experts, are crowd-sourced by

!"##$#%&'()*+&,-./0/1234&-05&6/784/0/1234&

&

&

&=&

users of the community in which they are used. While some scholars argue that tag clouds, which visualize the number of occurrences a tag has on a given site, confuse popularity with accuracy; others argue that allowing users to see what others are tagging a resource with can lead to more consistent tagging within a community, resulting in a tag cloud [that] translates the emergent vocabulary of a folksonomy into a social navigation tool (Sinclair and Cardew-Hall, 2007). Keshet (2011) explains the strengths of allowing user-implemented vocabulary in a retrieval system: Controlled vocabulary is typically used in a tree-like taxonomy to ensure effective indexing and to maintain the overall efficacy of the retrieval system by applying some degree of control to the indexing process. By controlling the indexing process using a so-called controlled vocabulary, index terms are standardized, and similar or related resources are collocated for ease of discovery by the user (Keshet, 2011). Folksonomies, in contrast, let the users define the categories, which also decentralizes the work of description and classification. The theory behind web 2.0 tagging is that a large community of users will create a collective knowledge environment, rather than rely on a few experts to dictate how things are defined and classified. Folksonomies represent a paradigmatic shift away from a definition of ontology that assumes a singular reality toward definitions that encompass diverse ways of producing, organizing, and utilizing knowledge (Keshet, 2011). Folksonomies create a faceted classification that mimics the Web itself. While hierarchical classifications are considered top-down, beginning with broad classes of knowledge and becoming more specific, folksonomies develop from the bottom-up, connecting related items through tags to form a larger body of resources (Taylor & Joudrey, p. 387, 2009). One of the benefits of faceted classification is that it allows folksonomies to grow in the areas that are of most interest to users. For example, Dewey could never have planned a place for the topic of iPhones within his classification system, and catalogers found a place for it under Generalities in the 000s. Goodreads, a folksonomy devoted to organizing books, now displays over 400 books related to the iPhone that have been tagged this way by Goodreads users. Unlike a taxonomy, which relies on an expert-created vocabulary, a folksonomy allows users to tag and search for items in the vernacular. For example, the word cookery is a Library of Congress subfield under the broader subject heading of food, though it isnt a word often used in conversational language. If a searcher using the social bookmarking website Delicious.com searched with the term cookery, it yields much fewer links (4,016 results) than its more commonly used synonym, cooking (744,049 results) (Delicious 2012). he use of commonplace language makes folksonomies more accessible for users who arent familiar with subject headings. Although taxonomies have a more structured format than folksonomies, folksonomy contributors have many reasons to create accurate tags. An article by Lu, Park, and Hu (2010) points out that taggers are motivated not only by a desire to organize their resources, but also by social incentives, namely: future retrieval; contribution and sharing; attracting attention; play and competition; self-presentation; and selfexpression (p. 764). These incentives encourage the users participation not just in tagging, but accurate tags that create a portrait of their interests and make for easy retrieval of information for themselves and others. Application of Taxonomies and Folksonomies Taxonomies Controlled vocabularies have been essential for decades in information retrieval:

!"##$#%&'()*+&,-./0/1234&-05&6/784/0/1234&

&

&

&>&

Without controlled vocabulary terms being assigned in MARC records, the keywords searched by users would not produce as many relevant results and as researchers tack on more keywords in their searches, they run the risk of search failure, or retrieving zero hits (Grey & Hurko, 2012). The reason for this is that controlled vocabularies are consistent and are arranged in taxonomy. Controlled vocabulary can also have many sub-headings or divisions within those categories in order to break down the results more precisely. Interestingly enough, subject headings add a lot to a keyword search. A 2005 study by Gross and Taylor found that if subject headings were to be removed from or no longer included in catalog records, users performing keyword searches would miss more than one third of the hits they currently retrieve (Grey & Hurko, 2012). Subject headings create links or relationships with other headings and help provide broader, related or narrower search terms, alike. However, taxonomies can only be used like this in controlled environments of information, like a library. They cannot be applied to searching on the web, for instance. There is simply too much information to catalog; it would be prohibitively expensive and it is ever changing. This is the main reason people started massing together to create collective tagging folksonomies: they wanted to organize content on the Internet. Folksonomies While folksonomies have only been around the last ten years or so, they are proving powerful in the digital age in a space where content is hard to classify and organize. The thing so many users like about folksonomies is that it is written in a natural language, using terms you would use every day. Terms can be changed and fall under several tags: Allowing library users to participate in describing library material by creating tags can provide a more appropriate description of the item specific to their peer groups. Scholarly research in continuously advancing subjects such as the sciences and engineering would greatly benefit from user-generated tagging because access of articles by those in the field with knowledge of specialized (Grey & Hurko, 2012). Folksonomies started off being used in the social networking world, such as Facebook and Twitter. They have spread to blog sites, photograph sites, news sites, etc. Sites like CNN.com use folksonomies to track trending stories and to group them into categories of similar interest. The future trend of folksonomies is to see them emerge in libraries, museums, and educational and corporate institutions. Peterson indicates that this idea has been met with some resistance: Folksonomies and tagging are being met with skepticism by some in the information sciences who argue that these schemes are philosophically relativistic and will lead to a system breakdown (2006). However, [o]ther information professionals appreciate the weaknesses inherent to folksonomies; yet still celebrate their potential for creative and dynamic information organization (Avery, 2010). The following section discusses the future of information retrieval in light of the strengths and weaknesses of both taxonomies and folksonomies. The Future of Search and Retrieval As discussed above, both taxonomies and folksonomies have their pros and cons in regards to user retrieval of information. The future of both methods is the subject of much of the current literature. These articles focus on folksonomies in particular, as this type of classification is newer and less integrated with systems used in libraries. It is clear that neither method is disappearing any time soon, although many scholars conclude that a retrieval system combining both methods could combine the strengths of both systems while compensating for the weaknesses of each.

!"##$#%&'()*+&,-./0/1234&-05&6/784/0/1234&
Taxonomies

&

&

&?&

Judging by the literature, the future of taxonomies appears stable. While it may seem that this method is outdated in comparison to dynamic, user-centered tagging that caters to digital information, it appears that traditional indexing remains the preferred method of information institutions. Because of the rules governing its creation and use, this type of taxonomical organization fulfilled users needs by being consistent and uniform, lending interoperability among the metadata schemas that use it (Matusiak, 2006). The Library of Congress Subject Headings (LCSH) and Medical Subject Headings (MeSH) are such established systems for cataloging, and Dublin Core and MODS use controlled vocabularies for assigning metadata to records. While each of these systems has its drawbacks, each functions as the basis for organization within communities that use it. The vocabulary used by each system remains the choice of professionals within information institutions. Some indexers view the subject headings they choose a conversation with users, leading to discussion on which terms should be used in a controlled vocabulary a point they likely considered in light of taxonomies, as this is what they allow the users to do (Matusiak, 2006). However, this is still a one-way conversation that does not involve user input. Social classification is not the answer to all information retrieval needs, but its use can lead to retrieval methods that are more user-centered. Indexing is performed in isolation from users. Especially for recently digitized items, social classification can introduce user language and views into a practice (indexing) that might contain antiquated terminology and processes (Matusiak, 2006). Folksonomies In many ways it is too early to tell what the future of folksonomies and tagging will be within information retrieval as its use is relatively new. A December 2006 Pew internet use survey indicated that 28% of internet users had tagged or categorized online content (Hayman and Lothian, 2007). This number had risen to 33% by 2008 (Pew, 2012). While more recent numbers are likely much higher than this, the userdriven organization of information is somewhat new compared to other retrieval techniques, resulting in a large amount of research on the subject. The usefulness of this technique varies by information need, and user preference. Thomas et al list three major strengths of user tagging of resources (Thomas, Caudle and Schmitz, 2010): Tags allow users to personalize their own resources Their tags supplement the controlled vocabulary with terms professionals may have left out The tags create user communities and the potential for discussion over resources within an information community

This last attribute is one that is focused on in much of the literature. Much of the literature indicates that tagging decreas[es] the cognitive distance between people working in an organization, which helps promote social capital, a key factor in knowledge sharing (Sinclair and Cardew-Hall, 2007). If there is a key advantage to including folksonomies in the future of information recall, it is the potential for dialog. Seeing how other people have tagged items can lead to new ideas for searches, which leads to a greater potential for resource discovery (Thomas, Caudle and Schmitz, 2009). This is important with the sheer volume of information particularly digital information available to be archived and organized. With millions of people organizing resources in millions of different ways, in short, things get messy: One criticism of folksonomies, however, is their uncontrolled nature and the inconsistency of their application. By definition, social tagging is carried out by untrained users in an uncontrolled environment. In particular, lexical anomalies (e.g. grammatical variations) and lack of synonym, homonym, and homograph control could result in a high probability of noise and recall in users search results at the expense of retrieval precision, making it difficult to pull together similar or related resources (Thomas, Caudle and Schmitz, 2010).

!"##$#%&'()*+&,-./0/1234&-05&6/784/0/1234&

&

&

&@&

However, messy does not necessarily mean ineffective. In fact, putting all resources into a single repository of miscellaneous items marked by tags can be more effective than sorting everything into categories as a taxonomy does. While subject headings place information into seemingly intuitive categories, users who have never located an indexed item before might not know where to begin looking. If their thought pattern does not follow the same track as the indexer, it might be difficult to find what they are looking for. Tagging allows users to dump all their information into one big, messy pile, and locate it later provided they take the time to tag it well before dumping it in with all the other information. At that point the item be it an article, photo, report, story has as many locations as it has tags. This allows the resource to be grouped with any other item it shares a tag with, revealing relationships that would have remained hidden in a taxonomy (Weinberger, 2007). An example of this is the photo-sharing website, Flickr. Flickr allows users to tag their own images, and add tags to others images as well, leading to better search results. This has proven to be extremely effective even though only single-word tags are allowed on the site, leading users to create unique tags linking words together that no searcher would likely use to find a photo to begin with (i.e. leaning-towerof-pisa, my*birthday, etc.) Still, user tagging lends itself to a powerful search on the site, with enough precision to distinguish between photos of Gerald Ford and Ford automobiles. Hayman and Lothian (2009) point out the strengths of a future that continues to employ this method: [I]it's useful when there's lots of information and the information is truly meaningful to individuals, it'll be adopted more and more widely. But we're also going to invent new ways to harvest tagging We'll also undoubtedly figure out how to intersect tags with social networks, so that the tags created by people we know and respect have more weight when we search for tagged items. In fact, by analyzing how various social groups use tags, we can do better at understanding how seemingly different worldviews map to one another (Hayman and Lothian, 2009). Combined Systems The consistency and established value of taxonomies have proven to be methods that institutions will continue to turn to in information organization, but the questions addressed by research into the dynamics of tagging have highlighted strengths of user collaboration that cannot be ignored. The strengths of both systems listed above and throughout this paper may be enough to convince readers that each system can independently fulfill users needs in the right context: The power of these hybrid approaches for information organization has been proven at such online commercial sites as Amazon, eBay, Flickr and non-commercial sites like Technorati, Wikipedia and Epinions. The ability to customize the routing, grouping and sharing of digital resources whether they be citations, images, datasets, or other digital formats combined with the ability to classify (catalog) using ad hoc vocabularies has the potential to greatly enhance the relevance of digital library content (Grey and Hurko, 2012). However, research has shown that in some cases one system on its own fails to address user needs. As demonstrated in various articles, controlled vocabularies and the tags that users choose to tag resources with often vary drastically. This demonstrates that users searching for articles and using the tags they deem intuitive would not find resources in a controlled vocabulary system if search terms were not adequately collocated in those institutions search functions. Thomas et al go on to highlight several cases where there has been serious divergence between professionally assigned subject headings and user-generated collective vocabularies (Thomas, Caudle and Schmitz, 2009): 2003: Nowick and Merring found that 40% of users terms did not match the 3 controlled vocabularies they compared them with 2006: Lin found only 11% overlap between users tags and Medical Subject Headings (MeSH) for the same documents

!"##$#%&'()*+&,-./0/1234&-05&6/784/0/1234&

&

&

&A&

2008: Wetterstrom found 75% of all tags examined did not match Library of Congress Subject Headings 2009: Rolla found that tags in Library Thing and brought out the same subject as did LCSH for the same resource, but used different words. He concluded that both the tags and the subject headings make useful additions to the librarys catalog/search function.

These findings indicate that while controlled vocabularies and the use of subject headings may be widespread, the terms used in these schemas do not always fit those that users find intuitive. Tags have the simple advantage of integrating users language (Thomas, Caudle and Schmitz, 2009). Consequently, systems that combine an established set of tags or subject headings could be substantially improved when supplemented by user-chosen descriptors (Grey and Hurko, 2012). While tags receive a considerable amount of scholarly attention as they become more and more the norm in online bookmarking or even social media sites, many authors believe that tags can work to strengthen metadata schemas such as Dublin Core and MODS, rather than replace them (Hammond et al, 2005). Examples of Combination Systems A variety of information professionals and institutions have come up with ideas for information retrieval that combine aspects of controlled vocabularies and tagging. These institutions, including education resources, public and academic libraries, and companies have attempted different ways to combine institutional and user-influenced retrieval methods. One such initiative that uses this approach, termed a folksonomy-directed taxonomy, is the Education Network Australia (edna), Australias primary online collection of resources for education and training. Indexers assign each resource subject headings using an education thesaurus. A current project called myedna allows users to personalize how the resources are displayed for them, but is also testing letting them add their own tags. Users can tag resources that are already tagged, making them more accessible to a wider range of users. Participants in the edna project identified some anticipated outcomes of the project that could apply to the rest of the information world. These, in short, include: Users can manage the resources with tags that are intuitive to the community in which it is relevant Consequently, a greater number of resources will be identified and shared by that community This collaboration could lead to more consistent categorization of resources within that community The community will form discussions and social networks around the tagging of resources, and generate better research and better organization (Hayman and Lothian,

Aside from Flickr, popular websites with different aims have also employed aspects of institutionally implemented keyword searching combined with user tags to make their products more findable. Amazon and eBay use this method so buyers can better search their products (Grey and Hurko, 2012). Both public and academic libraries have allowed users to contribute tags to their catalog as well. The University of Pennsylvania allows users to add tags to catalog records, and makes these tags available to other users. The Ann Arbor District Library project has a similar project, but did not note whether users were able to see other users tags (Thomas, Caudle, and Schmitz, 2009). Unstudied Implications of Social Classification As a final consideration, one aspect of the future of information retrieval sparks some interesting observations. The messiness of user-implemented tags discussed in the taxonomy section raises two considerations. The first was discussed above whether messy is really ineffective. The second causes us to wonder if tagging is accomplishing anything aside from information retrieval.

!"##$#%&'()*+&,-./0/1234&-05&6/784/0/1234&

&

&

&9B&

This paper has already discussed why users chose to tag things a certain way, which can lead to discussion on the development of better search and retrieval tools. The second question, however, has not been examined closely. Examination of tags on microblogging websites like Twitter and Tumblr indicate that there is also a social component to tagging that doesnt necessarily aid in information discovery. Literature and experience indicate that there are intentional and unintentional causes for this. Thomas et al found that around 40% of Flickr tags were misspelled or unidentifiable to dictionary software (Thomas, Caudle and Schmitz, 2010). One reason for this is likely that Flick allows only single-word tags, and users are forced to combine words into a single phrase. Tumblr allows multiple words to act as a single tag. Yet on both sites, users tag resources with phrases that no one would use as a logical keyword search to locate an item. Below is an example from a Tumblr blog:

Figure 2: Tagging on Tumblr No one would attempt to locate a picture of this breakfast using the keyword search ideal except for the peanut butter. This leads to two possibilities: either users are tagging items in this way because they are forced to, or they are doing it intentionally for reasons other than information discovery. It appears that this latter option may be true more often than scholars think. Users often tag resources in ways that do not increase its search results at all, but include information that solidifies users membership in the community in which they are sharing it (Hayman and Lothian, 2012). Another example from Tumblr illustrates this. The screenshot below shows tags a blogger has chosen for the music video to Jason Mrazs song 93 Million Miles. The user has tagged the song with the phrase I cant, which might not make sense to a non-Tumblr user. However, almost any regular Tumblr user will interpret this phrase to mean something like I can barely function because of the emotions this music gives me. While no one is going to search for a Jason Mraz music video using the phrase I cant, the tag is understood by users of the Tumblr community, and conveys information about the song to other users.

Figure 3: Tagging on Tumblr

!"##$#%&'()*+&,-./0/1234&-05&6/784/0/1234&
Conclusion

&

&

&99&

In conclusion, there are many things we still dont know about the future of search and retrieval in information science, but it seems safe to say that both taxonomies and folksonomies will be at the root of this practice. Years from now, these methods may not look the same as they do today the development of systems like the Library of Congress Subject Headings in taxonomies, and the transformation of folksonomies since Mosiac have demonstrated this. However, it is clear that, especially in an age of information overload, these two methods have a lot of potential to continue aiding users in fulfilling information needs for many years to come.

!"##$#%&'()*+&,-./0/1234&-05&6/784/0/1234&

&

&

&9:&

References

Avery, J. (2010). The democratization of metadata: Collective tagging, folksonomies and web 2.0. Library Student Journal, February 2010, Retrieved from http://www.librarystudentjournal.org/index.php/lsj/article/view/135/268. Cosentino, S. (2008). Folksonomies: The path to a better way?. Public libraries, 42. Delicious. (2012). Retrieved from http://delicious.com/search?p=cooking Facts about w3c. (2012). Retrieved from http://www.w3.org/Consortium/facts Fox, A. (2006). Cataloging for the masses. OCLC Systems & Services, 22(3), 166-172. Grey, A., & Hurko, C. R. (2012). So you think youre an expert: Keyword searching vs. controlled subject headings. Codex: The Journal of the Louisiana Chapter of the ACRL, 1(4), 21. Hammond, T., Hannay, T., Lund, B., & Scott, J. (2005). Social bookmarking tools (i): A general review. DLib Magazine, 11(4), 1-5. Hayman, S. & Lothian, N. (2007). Taxonomy directed folksonomies: Integrating user tagging and controlled vocabularies for Australian education networks. Paper presented at the World Library rd and Information Congress, 73 IFLA General Conference and Council Keshet, Y. (2011). Classification systems in the light of sociology of knowledge. Journal of Documentation, 67(1), 144 158. Kroski, E. (2009). The hive mind: Folksonomies and user-based tagging. InfoTangle. Lu, C., Park, J., & Hu, X. (2010). User tags versus expert-assigned subject terms: A comparison of librarything tags and library of congress subject headings. Journal of Information Science, 36(6), 763-779. Manktelo, M. (2010). History of taxonomy. Uppsala University Press, 1-2. Matusiak, K. (2006). Toward user-centered indexing in digital image collections. OCLC Systems & Services, 22(4). Miksa, F. (1984). The development of classification at the library of congress. University of Illinois Graduate School of Library and Information Science, (164). Peterson, E. (2006, November). Beneath the metadata: Some philosophical problems with folksonomy. D-Lib Magazine, 12(11). Retrieved from http://www.dlib.org/dlib/november06/peterson/11peterson.html.

!"##$#%&'()*+&,-./0/1234&-05&6/784/0/1234&

&

&

&9;&

Pew Internet Research. (2012). What internet users do online: Trend data (adults). Retrieved from http://pewinternet.org/Static-Pages/Trend-Data-(Adults).aspx Stone, A. (2000). The LCSH century: A brief history of the Library of Congress Subject Headings, and introduction to the centennial essays. Cataloging & Classification Quarterly, 29(1-2), Retrieved from http://www.catalogingandclassificationquarterly.com/ccq29nr1-2ed.htm. Sinclair, J. & Cardew-Hall, M. (2008). The folksonomy tag cloud: When is it useful? Journal of Information Science, 34(1). Retrieved from http://jis.sagepub.com/cgi/content/abstract/34/1/15 Taylor, A. G., & Joudrey, D. N. (2009). The organization of information. (p. 380). Westport, CT: Libraries Unlimited. Thomas, M., Caudle, D. & Schmitz, C. (2009). Trashy tags: Problematic tags in LibraryThing. New Library World, 111(5). Weinberger, D. (2007). Everything is miscellaneous: The power of the new digital disorder. New York, NY: Henry Holt and Company.

You might also like