| Social bookmarking sites generally | | | | classification of Internet resources |
| organize their content using tags. | | | | (such as web sites) is done by human |
| Social bookmarking sites are an | | | | beings, who understand the content of |
| increasingly popular way to locate, | | | | the resource, as opposed to software |
| classify, rank, and share Internet | | | | which algorithmically attempts to |
| resources through the practice of | | | | determine the meaning of a resource. |
| tagging and inferences drawn from | | | | This provides for semantically |
| grouping and analysis of tags. | | | | classified tags, which are hard to find |
| History | | | | with contemporary search engines. |
| The concept of shared online bookmarking | | | | Additionally, as people bookmark |
| dates back to April 1996 with the launch | | | | resources that they find useful, |
| of Within the next three years online | | | | resources that are of more use are |
| bookmark services became competitive, | | | | bookmarked by more users. Thus, such a |
| with venture-backed companies like | | | | system will "rank" a resource based on |
| Backflip, Blink, Clip2, Hotlinks, | | | | its perceived utility. This is arguably |
| Quiver, and others entering the market. | | | | a more useful metric for end users than |
| Lacking viable models for making money, | | | | other systems which rank resources based |
| most of this early generation of social | | | | on the number of external links pointing |
| bookmarking companies failed as the | | | | to it. |
| dot-com bubble burst. The contemporary | | | | Automatic Notification |
| concepts of social bookmarking and | | | | Since the classification and ranking of |
| tagging took root with the launch of the | | | | resources is a continuously evolving |
| web site oneview in 1999[1] and | | | | process, many social bookmarking |
| del.icio.us, in September of 2003. | | | | services allow users to subscribe to |
| Functional Overview | | | | syndication feeds (see RSS) based on |
| In a social bookmarking system, users | | | | tags, or collection of tag terms. This |
| store lists of Internet resources, which | | | | allows subscribers to become aware of |
| they find useful. Often, these lists are | | | | new resources for a given topic, as they |
| publicly accessible, and other people | | | | are noted, tagged, and classified by |
| with similar interests can view the | | | | other users. |
| links by category, tags, or even | | | | Disadvantages |
| randomly. Some social bookmarking | | | | There are drawbacks to such tag-based |
| systems allow for privacy on a | | | | systems as well: no standard set of |
| per-bookmark basis. | | | | keywords (also known as controlled |
| They also categorize their resources by | | | | vocabulary), no standard for the |
| the use of informally assigned, | | | | structure of such tags (e.g. singular |
| user-defined keywords or tags (see | | | | vs. plural, capitalization, etc.), |
| folksonomy). Most social bookmarking | | | | mistagging due to spelling errors, tags |
| services allow users to search for | | | | that can have more than one meaning, |
| bookmarks which are associated with | | | | unclear tags due to synonym/antonym |
| given "tags", and rank the resources by | | | | confusion, highly unorthodox and |
| the number of users which have | | | | "personalized" tag schemas from some |
| bookmarked them. Many social bookmarking | | | | users, and no mechanism for users to |
| services also have implemented | | | | indicate hierarchical relationships |
| algorithms to draw inferences from the | | | | between tags (e.g. a site might be |
| tag keywords that are assigned to | | | | labeled as both cheese and cheddar, with |
| resources by examining the clustering of | | | | no mechanism that might indicate that |
| particular keywords, and the relation of | | | | cheddar is a refinement or sub-class of |
| keywords to one another. | | | | cheese). |
| Advantages | | | | The separate (but related) tagging and |
| This system has several advantages over | | | | social bookmarking services are, |
| traditional automated resource location | | | | however, evolving rapidly, and these |
| and classification software, such as | | | | shortcomings could possibly be addressed |
| search engine spiders. All tag-based | | | | in the near future. |