User:BHoffman
From Open Annotation Collaboration
Brian Hoffman
Digital Library Publication and Access Manager
New York University Libraries
Use Case: Using the OAC Model for Annotation Interoperability
NYU Libraries publishes a wide-range of electronic material in support of scholarly research and communication. Some of our online publications result from collection oriented activities such as digitization, reformatting, or accessioning of born-digital materials. Others result from collaboration with communities of scholars seeking to develop new modes of scholarly communication that leverage the medium of the web. Through our development and stewardship of both sorts of publications, we have encountered a number of use cases or use-case domains pertaining to annotation.
For the purposes of this statement of interest, we draw a set of interrelated and interconnected use cases from the the MediaCommons Scholarly Network, a community-driven set of online experiments in digital publishing, 'peer to peer' review, social authorship, and non-conventional forms of scholarly output. MediaCommons is run on the open-source Drupal and Wordpress content management platforms, and is developed and supported by NYU Libraries' Digital Technology Services Group (DLTS). Brian Hoffman, Manager of Publications and Access for DLTS, will attend the workshop if invited.
MediaCommons has over 900 registered users, most of whom are media studies scholars from diverse academic and professional institutions. Collectively, the sites receive over 15,000 unique visits per month. The project receives institutional support from NYU Libraries and has been the recipient of an NEH Digital Humanities Start-Up grant. MediaCommons currently has three "projects" in production: MediaCommons Press; The New Everyday; and In Media Res. Two additional projects, #Alt-Ac and Open Canister, are currently being developed in collaboration with faculty and staff at the University of Virginia and Indiana University.
All of these projects involve regular publication of scholarly work for discussion, critique, or review. In each case, the underlying model of the blog (post, comment) has been modified or extended to support a novel form of publication and / or discourse. Some of these modifications and extensions are technical, including alterations to the basic data model that ships with Drupal or Wordpress; some are merely cosmetic, using CSS or other front-end devices to suggest more complex relationships between comments and their targets; and some are purely conventional: for instance, In Media Res is published on a "Theme Week" model, so that each day's post -- and the conversation that develops around it -- happens in the context of 4 other posts (and the conversations that develops around each of those.) The week's curator, the day's curator, and the day's conversation participants are all aware of this, and reference one another accordingly in the text of their comments (example: "Yes, this post and Ethanʼs go together quite well, giving us two very different perspectives on the production of web video..."). However, none of this is captured in the underlying data model, which only 'knows' of the relationships post:comment and comment:comment.
Through our work with the founding editors of MediaCommons and the various project editors, as well as other members of the community, we've noticed a number of more complex relationships that are implied in the way the community uses or wishes to use the sites. Though we have not collected a formal set of requirements, these kinds of relationships would include: comment-paragraph(s); comment-comment(s); commentpost(s); comment-conversation; and comment-region (video, image, etc.). Additionally, we have recognized that ultimately, there will be a need for relationships that are more layered and / or blurry. For example, 'This comment is aimed at Ethan's original post, but is also a rebuttal of both sides of the argument Aymar and Seth had last week', or 'This is in response to the conversation that took place on The New Everyday last month.'
Interestingly, DLTS attempted to take a step in this direction when developing the 1.0 version of The New Everyday in 2010. By making some minor extensions to the Drupal platform, we developed a commenting system that supported traditional comment-post relationships, but also allowed for secondary and tertiary targets, including other comments, posts, or an entire issue. Despite the relative simplicity involved in creating functionality to support this rudimentary step towards multi-targeted commenting, we withdrew the feature before releasing the site when we realized that the UI we had developed was too complex for the relatively low bar of entry we had set for site contributors. Thus, we feel that a major challenge for creating more nuanced relationships between annotations and targets will like in the interfaces or workflows that generate these relationships.
Another major challenge for our community of scholars will be developing annotations whose value goes beyond that of the comment. The MediaCommons managing editors are intensely interested in fostering community-driven 'peer to peer review'. That is, they envision a time when the discourse of scholars contributing and conversing on the sites will enable automated response to queries such as 'Who would be a good reviewer for my paper?'; 'How many revisions in published papers have been triggered by X's annotations?'; 'How valuable is Y's scholarship to this community?'. While we don't know to what degree this kind of intelligence will be possible, we feel that this aspiration makes this community a compelling case study for many diverse forms of social annotation.
