First some definitions:

• Social bookmarking is the process of bookmarking web resources, storing those bookmarks through an online service, and sharing them with other people.
• Tagging is a kind of free indexing, creating labels for a resource or it’s bookmarks that will help you and others find it. Bookmarks and tags are user created metadata.
• Folksonomies are essentially unstructured metadata schemas—groups of tags aggregated automatically because of some shared characteristic.

In his paper Ontology is Overrated, Clay Shirky makes an interesting argument about synonyms. He suggests that subtle differences in terms – like film, cinema, movies – actually communicate something about meaning that is lost when these terms are collapsed into a single controlled vocabulary term. Shirky argues people tagging “movies” have a different viewpoint about motion pictures than people tagging “films” or “cinema” and are more likely to find greater value in those items tagged by people who also prefer the term “movies.” In this instance, people with a broader interest, or those whose research aims require comprehensive results could simply search all three tags.

Integral to Shirky’s argument is an understanding of the scale of web resources. From a traditional classification standpoint, the ability to do an exhaustive search is important. All the resources on a subject in a library, or in a database need to be findable. But from a web standpoint this is rarely the goal. For many subjects the number of web resources is simply overwhelming and the goal is to find sufficiently relevant resources, not to find all resources. Precision is more important than recall.

Social bookmarking is not a new phenomena, it is almost as old as the Internet itself. According to Wikipedia, the first shared bookmarking service, itList, started in 1996. itList had all of the features of current social bookmarking applications like Delicious and Zotero.

The new wave of social bookmarking applications began with Delicious in 2003. “Delicious pioneered tagging and coined the term social bookmarking” (Wikipedia). In the years since, several applications have been developed. Some, like Delicious, serve a general web audience, others, like Connotea and Zotero are essentially citation managers aimed at an academic audience.

As we’ll see in the coming posts, libraries have engaged with both kinds of social bookmarking with varying levels of success.

It didn’t take long for me to come around to this point of view. The leg-work was already done in my first year chemistry class, when the Bohr model of the atom—with it’s tidy self-contained electrons happily circling nuclei—was tossed out in favour of electron clouds—mathematically defined regions where electrons may or may not be.

I fought against this for a while, this was the physical world after all, where things are or are not in a given place at a given time. But I lost. The world isn’t certain and fixed, right down to it’s tiniest particles, it’s probabilistic.

Once I’d come to terms with electron clouds, letting go of controlled vocabularies and classification schemes was a piece of cake.

Folksonomies are not heirarchical; there are no formalized relationships between the terms (broader-narrower, related, whole-part). They are probabilistic aggregates of tags, automatically grouped by computer program because they share some machine-readable characteristic (Adam Mathes). Whereas a traditional metadata schema is first written then applied to data sets, folksonomies work in the reverse.

The original folksonomies were popularity contests. And in most tag clouds you see  on blogs, on the sidebar at Delicious, on library homepages—what you are viewing are the most frequently used tags in the system. But it is possible to create more precise folksonomies. A tag search on Delicious generates a “related tags” cloud. Tags can also be aggregated by individual tagger, or by a group such as a Delicious network. I expect that new ways to combine and aggregate tags will evolve, allowing users to generate more precise folksonomies,  and to share them.

Peterson also quotes from Margaret Mann’s 1930 text Introduction to Cataloging and the Classification of Books:

“The cataloger must envisage the needs of the reader… He should, like the librarian, adopt a neutral stand between the reader and his books, giving emphasis to what the author intended to describe rather than to his own views.”

Neither of these statements about traditional cataloging is true of folksonomies. Because tagging classifies web resources, the rules of the physical world don’t apply. A can also be B. A can sometimes be B, and sometimes not, depending on the point of view of the person doing the tagging. And the point of view of the person doing the tagging absolutely does not have to be neutral. There is no attempt to try to create the artifice of objectivity in tagging. It is inherently a personal, biased, individual act.

It also a reader based act. Tags are applied to resources based on what the reader thinks they are about, or what the reader finds useful or interesting about them, not based on what the author intended. Peterson argues against folksonomies in part because of this reader bias.

“Because tags are relativized, personal, idiosyncratic views can coexist and thrive in the form of tags, in spite of their inconsistencies. Readers of texts on the Internet become individual interpreters, despite the document author’s intent.”

I disagree with her conclusion. I think an author can only control the content of the work, not how it is read, interpreted, understood, or valued by its readers. The creation of meaning in texts is collaborative.

These differences create a lot of discomfort in traditional cataloguing enthusiasts and are the source of some resistance to libraries completely embracing social classification schemes. If we had to choose only one system, I think we would stand to lose a great deal if we chose folksonomies. A great deal of very diligent, careful, and theoretically grounded thought has gone into the construction of top-down metadata schema and classification systems and in many circumstances they function much better than free tagging can ever hope to.

I think Peterson is correct when she claims: “folksonomies will not produce an efficient index.” As such, this would be a terrible way to construct a library catalog, or a database of academic journal articles. But when it comes to the web, if folksonomies will not produce and efficient index, who will? The scale and rate of growth of the web make cataloging it in any traditional sense impossible. Folksonomies add a layer of value. They add human input as an alternative to algorithms and web crawlers as a tool for locating items in the vast miscellany that is the web.

And libraries provide access to web resources. Tools like cloud tags from Delicious are one way libraries make use of folksonomies to enhance that access. Some libraries have even played with the idea of letting folksonomies add a layer of value to the more finite system of their library catalogs. Not an either/or approach but a both/and one.

“Delicious is a social bookmarking service that allows users to tag, save, manage and share web pages from a centralized source. With emphasis on the power of the community, Delicious greatly improves how people discover, remember and share on the Internet.”

Libraries provide access to web resources and many are incorporating Delicious tag clouds into their websites to improve that access. Tag clouds result from are a nifty bit of code that imports data from a bookmarking site like Delicious. The clouds are usually a block of text, alphabetically organized when read left to right, but with the font size of the tags varying based on the popularity of the tag. This creates a visually interesting and dynamic block of text.

Delicious goes public

Examples of libraries using tag clouds include Brimbank Public Library in Australia, and the Thunder Bay Public Library in Canada. Links to the homepage of each library can be found in the sidebar.

The Delicious tag cloud at both libraries is located with online resources – a logical place for it.

At Brimbank it’s found by clicking on eLibrary, then Weblinks. The cloud floats in the middle of the page and provides visual interest. I think the layout of the page and the size and prominence of the tag cloud would make it tempting for people to engage with it.

At Thunder Bay Public Library, the tag cloud is accessed by clicking on Virtual Collection, then Internet Links. On this website the tag cloud is less prominent. At the top of the page is a list of staff selected websites, and below that a “tag cloud” in list format. This placement of the cloud where you have to scroll down to see it makes it less likely people will interact with it. The format of the cloud as a vertical list, rather than a box of floating text, takes away some of the interest, though it may make the alphabetical organization of the tags easier to scan through.

It is possible on the Thunder Bay Public Library page to click a small “tag cloud”  link buried in a paragraph of explanatory text and retrieve a page which shows two tag clouds in more standard “cloud” format. On this page there are two tag clouds. One labeled TBPL’s Tags, and the other labeled Internet Links. Clicking on a tag in either cloud takes you to TBPL’s Delicious account, so the difference between the two clouds is unclear.

I think the model that Brimbank has adopted, of using the tag cloud to provide an immediate, visually interesting, and quick entry into their Delicious account, is the better choice. I suspect library patrons who are already Delicious users are unlikely to use these tag clouds—they would more likely just go straight to delicious to search for online resources. Delicious tag clouds on library websites serve to introduce new people to the concept of social bookmarking and how it can help them find online resources. To do this, a simple, obvious presentation of the tag cloud is likely to be much more effective than one where you have to dig to find it. If a patron does not know about Delicious or its value, what motivation do they have to dig through a site to find a tag cloud?

Delicious goes to college

Another interesting use of Delicious is happening in academic libraries, where it is being used to generate supplements to subject guides. The College of New Jersey Library subject guide for Italian Studies, for example, contains several sections generated by gathering tags from Delicious. The tags come from the libraries own Delicious account, so Delicious is really being used as a collocation and formatting tool. The library tags resources in Delicious, then uses Delicious feeds to gather and group appropriately tagged items in it’s subject guides.

Libraries have begun to interact with LibraryThing by making their catalogues available for browsing through the site. In additon, LibraryThing now offers LibraryThing for Libraries, which allows libraries to integrate library thing into their own catalogs. It works with any OPAC and consists of a few lines of HTML.

Libraries can chose to implement either or both of two options. The Reviews Enhancement package which allows libraries to enable patron reviews inside their catalogs, to upload existing reviews from LibraryThing, and includes blog widgets and a Facebook applications so that library patrons can promote their local library on their own social networking sites if they choose to.

The Catalogue Enhancement package includes book recommendations, enables tag browsing of your catalog, and enables easy linking to other editions and translations of a work. Whereas the Delicious tag clouds on library websites link uses out of the library page and into the Delicious site, LibraryThing tag clouds link users back into the library catalogue.

The Seattle Public LibraryThing

Seattle Public Library uses LibraryThing for Libraries. It isn’t possible to link directly to an item record in the catalogue, but it’s worth visiting the site and following along with my steps to see how LibraryThing is integrated into the catalog. In the search box on the home page, I searched for Harry Potter and the Deathly Hallows (assuming this title would definitely have tags). Then I chose the English language version of the book from the recall list.

At the level of the full catalogue record the integration with LibraryThing is seamless. There is a link to read or write reviews on the left. The holdings information is clear at the top of the page, and below that links to other editions held by the system, links to similar books held by the library, and a tag cloud of LibraryThing tags that take you to other books in the library that have been similarly tagged.

The main SOPAC page does have a list of links adjacent to the search window, but far fewer than are available through the tag cloud. At this stage, tags are not searchable, so the only way to gain the benefit from the tagging is through the links provided in the sidebar and the tag cloud. One of the drawbacks to this is that if you are looking for something that is tagged, but not in the top 500 tags, there is no way to use the tagging as an access point.

It is a free add-on for Firefox and is very easy to use. Once you’ve downloaded it, you have a little Zotero link in the bottom right hand corner of your web browser, and all you need to do to add a web resource is click it. Your Zotero pane opens in the bottom third of your browser and it automatically adds all the metadata it can gather from the resourse. You can edit this metadata, add tags, add notes, and attach word or pdf files so you have everything you need in one place.

When you set up your account, you are given the option to either make your research public or not. Zotero asks users to sign up for accounts as themselves, and warns them to think carefully about whether they want their research public or not. There is no way to keep some resources private, so if you opt to go public, everything is public. The only customization is that you can opt to make your resources public but keep your notes private.

Libraries that Zotero

Most academic libraries offer instruction in citation management products like RefWorks, and some are coming online with offerings in more social applications like Zotero. Emory University in Atlanta Georgia is one of these. From the Library homepage, clicking on guides, and then Zotero takes you to a great instructional page on this application.

In the left hand column is a listing of live workshops with dates and locations. The middle column contains a tutorial, with links to the Zotero site video tutorials. The right hand side brings a very friendly, Web 2.0 social feel to the page, with a photo and profile of the librarian responsible for the content, a chat box to interact with her when she is available, and contact information.

In addition, there is a feedback form to gather user feedback about how well the page satisfied their information need.

As more and more resources move online, as researchers become more independent, and better online tools are built to facilitate researcher-to-researcher communication, there is some question about what the role of research libraries will be going forward. I think librarians need to be well versed in these tools, and if they are able to provide high quality, timely instruction in these areas to undergraduate and graduate students and faculty, this will help enhance the profile of research libraries in the increasinging networked world.