Few days ago Hans Põldoja told me that Riina Vuorikari has been using the ecology concept in her social retrieval studies. It gives some input to the model of ecosystems (learning ecology). Specifically it shows that the reuse of traces that users leave to ecosystem is not so prevalent part of web-culture. Still, this behaviour may be used for cross-border translation between communities.
Ecology of social search for learning resources
Vourikari & Koper
In this paper they use attention metadata to model the ecology of social search.
An interesting part of the paper summarizes new search options:
Explicit search: comprises the traditional search box with text and filtering options based on multilingual metadata.
“Find by subject” offers browsing through pre-defined categories.
Personal search: Looking for bookmarks from one’s own personal collection of bookmarks
Novel exploratory search systems that assist users in obtaining content that meets their information needs include social navigation and collaborative recommender systems.
Social navigation involves using the behaviour of other people to help navigate online.
Social navigation types are: Interest indicators, which can be acquired either directly from the user (e.g. rating) or indirectly (e.g. time spent on an object).
Community browsing: these are social navigation features such as accessing resources through tagclouds and specific lists of most bookmarked resources, but also “pivot browsing” which means using tags or usernames as a means to reorient browsing.
Collaborative recommender systems use explicit ratings to find like-minded users (Adomavicius and Tuzhilin, 2005).
Rafaeli et al., (2005) introduced a system to harness the social perspectives in learning where the learner could choose from whom to take recommendations (friend or algorithm).
Koper (2005) used indirect social interaction in choosing a path that allows successful competition of a learning task.
Farzan and Brusilovsky (2005) studied social navigation and found that adding the time spent reading each page provides more precise insight into the intention of the group of users and more accurate information about pages selected from search results.
Secondly, they describe an interesting coordination system underneath the ecology of search that is based on three relations that might lead the people. I would think interpreting the personalized perception and actualization of such relations as search affordances.
Social bookmarking and tagging creates a triple (user, content, annotations) which indicates user’s relationship between resources, users, and tags (Golder and Huberman, 2006, Marlow et al., 2006, Sen et al., 2006). Such underlying structure allows flexible social navigation (e.g. tag-item, tag-user, user-item), but could also be a source for collaborative recommender systems by linking like-minded users not only through resources, but also through tag-based interest sharing (Santos-Neto et al., 2009).
The use of social information (annotations – interest indicators) in navigation was smaller than i expected, still the traditional search culture prevails. However, social search as a way of using community traces appeared.
A model was created of data to study how processes are interlinked (i.e. ecology).
The model shows that the annotation (i.e. Interest indicators) play an integral part in creating a social search ecology and offer more diverse ways to discover resources.
The search taking advantage of Social Information Retrieval methods yield more relevant resources with less effort from the user. Despite this edge, users have a strong search preference for Explicit search methods (2/3 of all executed searches).
Most often users discover cross-boundary learning resources as a result of Explicit search, and when the resource is deemed relevant, they bookmark it in the Search result list.
Conclusion: We show that users are more efficient with Social Information Retrieval
strategies, however, Community browsing alone does not help users discover a wider variety of cross-boundary resources.