Hypothesis hosted the second I Annotate meeting in San Francisco from April 3-6, and we think it was a tremendous success.
I Annotate brings together the worldwide annotation community to share developments, insights, challenges, and forge new connections. This year there was a marked sense of progress from last year, with a noticeable move from often early-stage pilots or standalone applications to a growing sense of the possibility of a unified annotation fabric based on open standards. Discussion of the hard problems associated with moving annotation into the mainstream abounded. Many talks are now available in presentation format on the agenda page.
The preface to I Annotate was set by a meeting organized by the W3C [video] to consider the charter for a formal Working Group on annotation standards. The draft charter is undergoing revision, but it was obvious that there was broad consensus on the overall aims for the group, which would be chartered for a period of two years initially. The W3 will play a critical role in providing a unitary framework within which an annotation data standard can be discussed, along with serializations and API specifications. Ivan Herman, the head of the W3C’s digital publishing group, observed that his organization’s interest in annotation arose out of the growing convergence of interest in annotation across many different communities.
A wide range of talks highlighted the many early successes of annotation. An energetic presentation by Tom Lehman, one of the founders of Rap Genius [video], which has utilized a proprietary approach to annotation, focused the audience on the challenge of encouraging and establishing a wide user base. Lehman encouraged the audience to not get too hung up on technical arcana, but instead excite users with a simple and easy to use interface. In turn, this led to later discussions on the apparent hegemony of the browser right-pane interface for annotation editing, and the lack of a compelling UX for mobile annotation authoring.
Phil Desenne of Harvard University spoke about the integration of the Annotator open software library into edX’s online education platform at Harvard [video], and its impact the first trials. In the pilot of these tools, an online class on the poetry of Walt Whitman, over 1100 students (about 12 percent of those enrolled) made 60,000 annotations, with an average of 52 annotations per user. The most popular target of annotations was the poem “Song of Myself,” which drew over 22,000 comments. HarvardX’ success in this project highlights the need for a responsive UI to accommodate thousands of annotations over small sections of texts, and a robust way of linking passages across multiple annotation targets.
Eric Weisstein of Mathematica discussed the need for a worldwide digital library of mathematics which could be assembled through the annotation of mathematical formulae discovered online. A sample project to discover continued fractions, a form of numerical representation, unearthed 1,300 core and 11,000 derived fractions, semantically annotated to facilitate natural language search. Mathematica’s goal is to build an infrastructure that allows content experts to mark up and annotate mathematical results, helping to build a digital mathematics library.
Working in a historical framework, Sean Boisen of Logos software demonstrated a similar need for semantically enriched annotation [video]. The Bible is possibly the single most heavily annotated textual work in history, with a Judeo-Christian legacy originating at least from 2nd Century CE Judaic commentary of the Torah. One of the most amusing historical biblical-era annotations is found in an early Greek text from the 3rd Century CE, where a reader wrote alongside the text to the editor, “Fool and knave, leave the old reading and do not change it.” Logos has enabled cross-language association of terms and concepts, including both personages and geographic place names. In addition, Logos has added contextual enhancement so that referants for pronouns and descriptive phrases can be associated with speakers and places, such that a phrase “He said …” could be discovered by a user in a search for “Paul.”
V. David Zvenyach, a lawyer for the District of Columbia, eloquently spoke of the circuitous path that legislation takes as it is translated into legal code [video], with each step involving a different document, and annotations appearing at several stages. These annotations are difficult to incorporate, and difficult to use. Zvenyach has implemented a draft system for tracking and annotating legal code for the District using GitHub, the software revision control system. Ideally, the public should be able to issue pull requests from the Git repository and make their own annotations, proposing changes in legal code or draft legislation. Enabling a simpler interface to Git, and the integration of pull requests into an annotation system, is a service that Zvenyach feels would ultimately help make better law.
The hackdays were a great opportunity for the many working with open annotation to get together. Over 40 people turned out for two days of hacking and discussion. We saw new programmers getting their feet wet with Python and having success running their own annotation servers, enabling us to ensure our install procedures are well documented. The majority of the teams worked on various improvements to Annotator 2.0 [video], including work on browser support (i.e. getting things to work in Internet Explorer), documentation, planning release schedules, moving everything to a new open annotation GitHub repo, discussing new features, launching the new annotatorjs.org website, and more. Overall, it was a wonderful opportunity for the community to meet face to face.
We thank the Alfred P. Sloan Foundation for the funding which made this meeting possible.