Position Paper for Annotation Workshop

Frederick Hirsch, Vlad Stirbu; Nokia

W3C Workshop on Annotations
2 April 2014, San Francisco, CA, USA

Interest in Participating in Workshop

The Web and Internet Technologies team in the Nokia CTO office has been actively involved in the Open Web Platform for many years, including chairing at the W3C. We have also been working on applying the Open Web Platform to education and recognize the issues and importance related to publishing and educational applications. Annotations are an important part of the technology platform with interoperable, open and clearly defined formats and meaning being essential. The Open Annotation Data Model Community Draft and Annotation Use Cases draft produced by the W3C Open Annotation Community Group are excellent contributions to the topic and clearly there is much more that can and should be done, as discussed on the Open Annotation Community group mailing list. Using this workshop to better understand the area of work that might progress in a W3C Open Annotation Working Group (draft charter) seems appropriate at this time and we are interested in being part of that discussion.

Use Cases

We have identified two use cases that are immediately applicable to the educational scenarios we have been pursuing. Open annotation is core to these cases.

Student Annotation of Course Materials

The first use case is for students to be able to make notes by annotating the materials associated with their course. This includes the ability to add, view and search annotations as well as sharing with others. Access controls on annotations are essential, enabling a student to keep them private, share within a study group, share with a teacher, or make more widely available. Search of annotations is a topic that may require additional metadata in addition to the ability to search text (for textual annotations).

Annotations for Assignments

The second use case consists of students completing assignments by creating annotations. Annotations can be responses to items in an assignment sheet, in a book or elsewhere. The ability of both annotations and targets of annotations to be a variety of types (e.g. text, video, images etc) and/or portions of types as suggested in the Open Annotation Data Model is very useful. For example, students could annotate portions of a film to add commentary on scenes as part of project to analyze the film.

We expect students would be able to save their annotations to a server, and work on them from a variety of their devices at different times (e.g. at home, school or while mobile). Once a student completes an assignment by creating a number of annotations, grouped into a collection, this collection could be shared with the teacher (it would likely be an ordered collection to ease processing). The use case goes on to suggest that the teacher might have comments on some of the responses, and thus create annotations on some annotations (or the assignment set as a whole) and then return it to the student for further work. In this case the student might respond to teacher annotations with additional annotations, add annotations on the assignment itself (increasing the size of the annotation collection) or even modify previous annotations. Once this is done the student might share the revised collection with the teacher.

This scenario suggests the need for access control as well as metadata associated with annotations, including the author, the time of the annotation and perhaps even a version history. Thus annotation provenance is essential from the start. Clarity and the ability to provide provenance information on all aspects of the model is important as outlined in the example given on the Open Annotation Community Group mail list.

Common Requirements

In addition, we expect to store annotations on a server, so a simple, concise, portable and web-friendly annotation serialization would be useful. The JSON serialization using the JSON-LD context to simplify the representation while also permitting RDF usage for those that need it seems appropriate and useful. This was noted on the Community Group mailing list (2nd example).


Annotations will be important to our use cases for educational use of the open web platform, both in the context of browser access to educational materials as well as the use of ePub materials. Open and interoperable annotation formats will be necessary for sharing annotations among systems and users. The Open Annotation Data Model offers a flexible and useful foundation and JSON-LD looks like it enables a means for sharing annotations with a simple serialization, hiding much underlying complexity. A REST API for creating, retrieving and managing annotations should be useful to our use cases. A client Javascript API might also be useful, for example enabling client-side checking that an assignment is complete before submitting to a teacher. All of these topics warrant discussion at the workshop.

We expect that the workshop will also help inform the chartering of a working group at the W3C. We look forward to seeing participation from browser vendors, ePub experts and end users. We anticipate that the workshop discussion will include a review of the draft charter, discussion of priorities and how to manage scope to achieve success. One item that is mentioned in our use cases is the ability to search annotations, so we suggest mentioning that explicitly in the charter in the context of the REST interface. We also question whether an abstract model needs to be developed in the working group given the existence of the Community Group 'Open Annotation Data Model'. Perhaps the focus should be bringing that document quickly to Recommendation status.

The Data Model community draft from the Community Group gives the initiative a good start on a solid foundation. The workshop is timely and important and should help raise issues, clarify scope and direction. We are interested in this discussion which is relevant to our project.