Converted from a Word document
As a community of practice, digital humanists deal with data and metadata not as static artifacts, but rather as complex, multi-dimensional and multi-layered datasets that can be analyzed, annotated and manipulated in order to produce new knowledge. One of the most important challenges facing DH today is how to consolidate and repurpose available tools; how to create reusable but flexible workflows; and, ultimately, how to integrate and disseminate knowledge, instead of merely capturing it and encapsulating it. This technical and intellectual shift can be seen as the “infrastructural turn” in digital humanities (Tasovac et al. 2015).
Application Programming Interfaces (APIs) have the potential to be powerful, practical building blocks of digital humanities infrastructures.
On the technical level, they let heterogeneous agents dynamically access and reuse the same sets of data and standardized workflows. On the social level, they help overcome the problem of “shy data”, i.e. data you can “meet in public places but you can’t take home with you” (Cooper 2010). Some 10 years ago, Dan Cohen started the conversation about APIs in DH by pointing out that, despite their potential, Andreas few humanities projects — in contrast to those in the sciences and commercial realms — were developing APIs for their resources and tools (Cohen 2005). In the decade since, API development in the digital humanities has certainly increased: today, both large-scale, national and international initiatives, such as HathiTrust, DPLA or Europeana, as well as individual projects, such as Canonical Text Services (CTS), Open Siddur, Folger Digital Texts, correspSearch etc., are focusing their attention and resources on developing APIs. It is now time to reflect on this development: have standards or best-practices evolved? What workflows are most effective and efficient for creating APIs? What are the challenges or stumbling blocks for creating or using APIs? Are APIs being used by DH researchers? What is the future of API development and use in the humanities community?
Infrastructures are installations and services that function as “mediating interfaces” or “structures ‘in between’ that allow things, people and signs to travel across space by means of more or less standardized paths and protocols for conversion or translation” (Badenoch and Fickers 2010, 11). Digital research infrastructures are no different: they are a mediating set of technologies for research and resource discovery, collaboration, sharing and dissemination of scholarly output.
This panel will cover both the theory and practice of APIs in the digital humanities today. It will bring together researchers working on major European and North American projects, who will discuss APIs from the perspectives of design, implementation, and use, as well as technical and social challenges. Each group will have 10 minutes for their statement, and 40 minutes will remain for group discussion and questions from the audience. One of the panel members will serve as the moderator. All speakers have confirmed their intention to participate in the panel.
Toma Tasovac (Belgrade Centre for Digital Humanities) will discuss an API-centric approach to designing and implementing digital editions. Starting with the notion of text-as-service and textual resources as dynamic components in a virtual knowledge space, Tasovac will show how two recent projects —
Raskovnik: A Serbian Dictionary Platform and
Izdanak: A Platform for Digital Editions of Serbian Texts — were implemented using API-focused data modeling at the core of the project design process. The API-first approach to creating TEI-encoded digital editions оffers tangible interfaces to textual data that can be used in tailor-made workflows by humanities researchers and other users, well-suited to distant reading techniques, statistical analysis and computer-assisted semantic annotation. The “infrastructural turn” in Digital Humanities does not only have practical implications for the way we build tools and create resources, but also has theoretical ramifications for the way we distinguish highly from loosely structured data: if text is not an object, but a service; and not a static entity, but an interactive method with clearly and uniquely addressable components, a formal distinction between a dictionary and, say, a novel or a poem, is more difficult to maintain.
Clifford Wulfman and
Natalia Ermolaev (Center for Digital Humanities, Princeton) will discuss the design and implementation of Blue Mountain Springs, the API for the Blue Mountain Project’s collection of historic avant-garde periodicals.
By modeling magazine data using the FRBRoo ontology and its periodical-oriented extension PRESSoo (PRESSoo, 2014), this RESTful API exposes the Blue Mountain resource in a variety of data formats (structured metadata, full-text, image, linked data). The authors will provide several examples of how Blue Mountain Springs has been used by researchers, drawing especially from the results of the hackathon they will host at Princeton in February 2016, which will bring together approximately twenty periodical studies scholars, technologists, and librarians to work with the API. Creating APIs is part of a trend in DH to move into a post-digital-library phase, when the traditional library functions of discovery and access are no longer sufficient to support research in the humanities. This trend also suggests that DH researchers must reconceptualize their own engagement with material, to think less in terms of monographs and more in terms of resources, and consequently to promulgate their work not as web sites but as web services.
http://bluemountain.princeton.edu
Traditional scholars have been citing texts the same way for centuries : for example, “Hom. Il. 1.1”, which corresponds to Homer’s Iliad, Book 1 Line 1. However, although the passage identifier does not change from one scholar to another in most cases, the abbreviation used for the author and the work title will diverge among authors, countries, and publications. Thibault Clérice (University of Leipzig) will discuss the design of the Canonical Text Services (CTS) and its URN scheme, which make the traditional citation system used by classicists machine-actionable (Blackwell and Smith 2014)
http://www.oeaw.ac.at/icltt/node/193Adrien Barbaresi from the Austrian Academy of Sciences (ICLTT) will discuss the use of APIs in building resources for linguistic studies. The first case deals with lesser-known social networks (Barbaresi 2013) while the second tackles the role of the Twitter API in building the ICLTT’s "tweets made in Austria" corpus
Finally,
This is a result of the confluence of several factors: that the research data most humanists want to access is seldom available via an API; that many sources that do offer relevant data lack the structure or metadata to make the API useful; and that humanistic researcher generally lack the skill set to experiment directly with APIs. At the same time, however, one of the key desires expressed by researchers in the current landscape is the federation of high-quality data (Edmond and Garnett 2015). Jennifer Edmond and
Vicky Garnett (Trinity College Dublin), will provide reflections on the place of APIs within European research infrastructures for the humanities. Their contribution to the panel builds on their recent study on the Europeana Cloud project, which found that while access to data is a real and growing area of interest, very few humanities researchers seem to actively and directly use APIs.
The five speakers on this panel will address some of the most pressing issues related to the ongoing development and future of APIs on the DH research infrastructure landscape. The discussion will cover both micro- and macro levels, ranging from methodological implications and technical scalability to the ways in which API-based data access to collections challenges traditional norms of institutional identity and independence. As such, the panel will offer a timely platform for a multifaceted debate on the potentials and pitfalls of building and using APIs in the digital humanities.