Phase II Executive Summary

Annotating is a practice core to scholarship. Marginalia, glosses, and other recognized forms of annotation left behind by authors and by previous readers of a text bring an added dimension to how we understand that text, sometimes serendipitously, often through intent. The experience of reading can be altered by the presence of annotations. A student reading a well-annotated text will come away with a deeper understanding and appreciation of that text than he or she had before. Annotation of other media or in other forms can be just as powerful. In their survey of scholarly annotation practice past and present, Maristella Agosti, et al. note that annotation is often in its own right a "vehicle for carrying and transmitting ideas and knowledge to other people" and has been used historically as a "professorial tool." The potential formats and modes of annotation should expand in a digital context. In his 2000 presentation at King's College London, John Unsworth identified annotating as a "scholarly primitive," but noted with concern the lack of progress towards applications that effectively supported the sharing of annotations on the Web. This lack persists today. A number of issues and perceived obstacles have slowed development and adoption of tools for scholarly annotation of digital resources. Some of these have to do with the difficulties of mapping traditional practices of annotation into a digital context, but many others reflect a failure to systematically exploit opportunities afforded by the semantic Web and linked data initiatives, an unevenness in tools and limited options for annotating across clients and content repositories. Missing is implementation of a shared data model and ontology on which to base interoperable annotation tools and services.

The OAC seeks to remove these obstacles by engendering the emergence of a Web and Resourcecentric interoperable annotation environment and by demonstrating the potential of this environment to support robust scholarly annotation in a multiplicity of digital humanities and digital library domains and across boundaries of clients, tools, and repositories. OAC Phase I focused on the development of a foundational data model and ontology for interoperable scholarly annotation, leveraging the collective experience of OAC members with Web architecture, model-theoretic semantics, Linked Data Initiative principles, and existing and emerging Web standards and best practices. OAC data model development has proceeded to the point where it is now feasible to broaden testing and, in collaboration with scholars and repository managers, to vet and refine the OAC data model against real-world scholarly use cases involving significant repositories of digital content. Work carried out during OAC Phase I has made clear that adoption of a shared annotation data model suitable to support humanities scholarship must happen initially domain by domain, and must begin happening soon to avoid the de facto entrenchment of a tower of Babel, project-by-project approach to scholarly annotation on the Web. To further our understanding of annotation interoperability, to create the Web and Resource-centric interoperable annotation environment sought, and to encourage in support of these goals the adoption and evolution of the open and robust data model for scholarly annotation on the Web developed during OAC Phase I, we will need to focus in OAC Phase II on directly engaging humanities scholars and involving existing collections of digital content having well-defined communities of scholars interested in annotating such content.

Accordingly, the continuing members of the OAC, i.e., the universities of Illinois, Maryland, and Queensland and the Los Alamos National Laboratory, request a total of $673,942. to undertake OAC Phase II in order to execute concrete, collaborative small-scale demonstration projects and experimentation featuring challenging scholarly use cases, existing digital content in situ, and the direct involvement of scholars and curators of scholarly digital resources. Through this demonstration and experimentation we will vet and refine the nascent OAC data model and ontology, demonstrate its utility in enabling a resource centric environment for interoperable scholarly annotation on the Web, and proactively encourage early adoption of this environment across a range of different scholarly domains. We propose to implement a total of 8 complementary annotation demonstration experiments, each involving active projects external to OAC that already have existing digital content and well defined scholarly audiences. Use cases will span a range of media formats, will address a complementary range of modeling requirements (e.g., a range of different approaches to segment description), and will include communities with prototype annotation tools for scholarly annotation or at least well-recognized and understood annotation tool needs. We propose to configure each demonstration to be a joint undertaking involving an OAC member institution team working in close collaboration with the repository staff and scholars proposing the use case.

Drawing on this experimental work, we also in OAC Phase II will advance the maturity, completeness and expressiveness of the OAC data model itself, the guide to this data model, and the ancillary documentation and best practice recommendations intended to inform implementers and facilitate consistent use of the data model by a broad range of scholarly communities. The latter category of documents will include recommendations for serializing Annotation Resource Maps in RDF, RDFa, and other formats. An outcome of Phase II will be a version 1.0, suitable-for-production OAC data model and specifications of similar maturity and completeness to 1.0 specifications developed for earlier projects such as OAI-PMH and OAI-ORE. Though not an ultimate end in itself, an interoperable data model of and ontology for scholarly annotation on the Web is a necessary pre-requisite to enable our larger goals.

Finally, we will observe closely the annotation demonstration projects in progress and require that all 8 projects collaborating with us during OAC Phase II report back in depth on their experiences along the way in applying the OAC data model to meet their domain-specific needs. This documentation in combination with what we learn by observation from each demonstration about needs and practices with regard to scholarly annotation of digital resources will be synthesized as part of our own report out. By documenting and synthesizing the results, both positive and negative, from the 8 demonstration experiment projects, we will be able to advance the state of our understanding of scholarly annotation needs in the Web environment and provide an adaptable and well illustrated model for others interested in implementing the OAC data model and ontology. There is broad-based interest in this problem space, and it is our expectation (depending in part on the success of our 8 demonstration experiments) that many other researchers and funders within multiple other domains will benefit from and be interested in further developing the OAC Web and Resource-centric interoperable annotation environment that we will engender and demonstrate through the work proposed for OAC Phase II.

See the Phase II Proposal as presented to the Andrew W. Mellon Foundation in October of 2010.