Digital Scholary Editions as Interfaces -  - ebook
Opis

Interfaces are important elements of digital scholarly editions as they allow and direct the interaction of users with the online content and they facilitate the access to and exchange of data and information. Some interfaces are created for the human user (GUI), others for machine interaction and data exchange (API). Both aspects of interfaces and their roles in digital scholarly editing were discussed at a conference in 2016 organised by the Centre for Information Modelling at the University of Graz and the Digital Scholarly Editions Initial Training Network DiXiT. This volume includes a range of papers presented at the conference that highlight the diverse views and approaches towards interfaces in the digital scholarly editing community.

Ebooka przeczytasz w aplikacjach Legimi na:

Androidzie
iOS
czytnikach certyfikowanych
przez Legimi
Windows
10
Windows
Phone

Liczba stron: 590

Odsłuch ebooka (TTS) dostepny w abonamencie „ebooki+audiobooki bez limitu” w aplikacjach Legimi na:

Androidzie
iOS

Contents

Preface

Introduction

Roman Bleier, Helmut W. Klug

Discussing Interfaces in Digital Scholarly Editing

Theorizing the Interface

Tara L. Andrews, Joris J. van Zundert

What Are You Trying to Say? The Interface as an Integral Element of Argument

Wout Dillen

The Editor in the Interface: Guiding the User through Texts and Images

Shane A. McGarry

Bridging the Gap: Exploring Interaction Metaphors to Facilitate Alternative Reading Modalities in Digital Scholarly Editions

Ginestra Ferraro, Anna-Maria Sichani

Design as Part of the Plan: Introducing Agile Methodology in Digital Editing Projects

The Interface in Practice

Stefan Dumont

Interfaces in Digital Scholarly Editions of Letters

Chiara Di Pietro, Roberto Rosselli Del Turco

Between Innovation and Conservation: The Narrow Path of User Interface Design for Digital Scholarly Editions

Joshua Schäuble, with Hans Walter Gabler

Encodings and Visualisations of Text Processes across Document Borders.

Elli Bleeker, Aodhán Kelly

Interfacing Literary Genesis

Jeffrey C. Witt

Digital Scholarly Editions and API Consuming Applications

Hugh A. Cayless

Critical Editions and the Data Model as Interface

Evaluating the Interface

Federico Caria, Brigitte Mathiak

A Hybrid Focus Group for the Evaluation of Digital Scholarly Editions of Literary Authors

Elina Leblanc

Design of a Digital Library Interface from a User Perspective, and its Consequences for the Design of Digital Scholarly Editions: Findings of the

Fonte Gaia

Questionnaire

Appendix

Biographical Notes

Preface

The present volume “Digital Scholarly Editions as Interfaces” is the follow-up publication of the same-titled symposium that was held in 2016 at the University of Graz and the twelfth volume of the publication series of the Institute for Documentology and Scholarly Editing (IDE). It is the result of a successful collaboration between members of the Centre for Information Modelling at the University of Graz, the Digital Scholarly Editions Initial Training Network DiXiT1, a EC Marie Skłodowska-Curie Action, and the IDE. All articles have undergone a peer reviewing process and are published in Open Access. They document the current state of research on design, application and implications of both user and machine interfaces in the context of digital scholarly editions.

The editors of the volume are grateful to the Marie Skłodowska-Curie Actions for enabling not only the symposium in 2016 but also the publication of the present volume with their financial support. Special thanks are also due to the staff of the Centre for Information Modelling, above all Georg Vogeler, who contributed to the successful organisation and completion of the symposium and this volume with their ideas and continuous support. Furthermore we want to thank all authors as well as all peer reviewers for the professional cooperation during the publication process. Last but not least we want to thank the many people involved in creating the present volume: Barbara Bollig (Trier) for language corrections and formal suggestions, Bernhard Assmann and Patrick Sahle (Cologne) for support and advises during the typesetting process, Selina Galka (Graz) for verifying and archiving (archive.org) all referenced URLs in January 2018, Julia Sorouri (Cologne) for the design of the cover as well as the artist Franz Konrad (Graz), who provided his painting “Desktop” (www.franzkonrad.com/gallery/desktop-2008-2010/) as cover image.

We hope you enjoy reading and get as much intrigued by the topic “Digital Scholarly Editions as Interfaces” as we did.

Graz and Berlin, September 2018, the editors

1dixit.uni-koeln.de

Introduction

Discussing Interfaces in Digital Scholarly Editing

Roman Bleier, Helmut W. Klug

University of Graz

Abstract

Digital Scholarly Editions as Interfaces, edited by Roman Bleier, Martina Bürgermeister, Helmut W. Klug, Frederike Neuber, Gerlinde Schneider. Schriften des Instituts für Dokumentologie und Editorik 12. Books on Demand, 2018, V–XV.

Interfaces define how research material is presented. They shape the view recipients acquire from historical sources. Since the digital medium is more open to variations than the once traditional form of presenting Scholarly Editions in printed book form, discussions on how to deal with the new possibilities started at a very early stage after the emergence of digital scholarly editions. In the beginning these were strongly influenced by traditional presentation practices but have shifted to aspects more associated with the digital paradigm. Theoretical approaches towards interfaces, however, were only sporadically published and have been continuously demanded by the scholarly community. This introduction attempts to summarize the scholarly discussions on interfaces and provides an overview of the papers presented in these proceedings: they offer both theoretical approaches and discussions of practical implementations together with studies evaluating interfaces.

Early 2016, a heated discussion sparked off at the Centre for Information Modelling at the University of Graz about the role of interfaces in digital scholarly editions (DSE) and the question of whether the DSE itself takes up the role of an interface between documents, users and machines. This discussion led to the decision to hold a conference about the topic in September of the same year - entitled “Digital Scholarly Editions as Interfaces”. The aspired format for the conference was a moderate setting to provide a stage for early career researchers and the fellows of the Digital Scholarly Editions Initial Training Network DiXiT, a European Commission Marie Sklodowska-Curie Action, to present their projects, ideas and ongoing research on the relationship between interface and DSE. The overwhelming response to the call for papers revealed the strong interest in this topic of the Digital Humanities (DH) community. The result was a densely packed two day symposium with an international audience and speakers. Accounts of the event can be found on the DiXiT blog (Bleier et al., “Report”) and on H-Soz-Kult (Bleier et al., “Tagungsbericht”).

The conference programme was framed by two inspiring keynote presentations by Dot Porter of the Schoenberg Institute for Manuscript Studies at the University of Pennsylvania and Stan Ruecker of the School of Art and Design at the College of Fine and Applied Arts at the University of Illinois. Porter opened the conference. In her presentation “What is an edition anyway?” she discussed different definitions of edition and revealed the results of her latest survey on the usage of editions (printed, digitized, and digital). She related the survey results to her earlier surveys (from 2002 and 2011) and showed that a certain inhibition threshold exists towards using the digital medium as means of presentation. Her bold closing statement “Data over Interface” provided plenty of grounds for lively discussions throughout the conference whether the data or the interface are the integral part of a DSE, and in consequence, whether the machine or the human user is the primary addressee of a DSE.

In contrast, Stan Ruecker set the focus of his keynote on the aspect of the design process in general and the design of creative and experimental interfaces for Humanities and Cultural Heritage purposes in particular. He emphasized his arguments with examples from various research and design projects strongly focusing on the experience of scholarly readers. He concluded that the social and dynamic aspect of scholarly reading should play a more important role in the designing of DSEs.

According to the Oxford English Dictionary interfaces are:

A means or place of interaction between two systems, organizations, etc.; a meeting-point or common ground between two parties, systems, or disciplines; also, interaction, liaison, dialogue.

This rather general definition still hints at the complexities the concept of “the interface” incorporates. In DH it not only contrasts humanities with computer sciences and design aspects but also human to human with machine to machine communication as well as human computer interaction. The role of the interface and along with this role, its definition varies according to the person, domain, interest, use (etc.) it is associated with. Interface research deals with a vast complexity of the research topic, the amount and particular structure of humanities’ data, as well as the diverse collection of affected research domains. Equally demanding are the rapidly evolving technology, with the ever changing demands on both publishers and recipients of editions and interfaces. Additionally, over hundreds of years the distribution of knowledge has been associated with the form and feel of the book (Burdick et al. 139). The Scholarly Edition has a long tradition in print and, hence, these experiences and associations with the printed edition strongly shape the way how DSEs are designed today. However, is it indeed necessary for editors to use a skeuomorphic design approach, trying to mimic the book? Are only DSEs that follow the in order to accepted and used by the scholarly community? To what extend does such an approach hinder innovation and the development of more efficient solutions for DSEs? (Pierazzo 170–175)

Understanding the DSE itself as an interface means understanding it as a connection point between historical documents and the user, whether a human being or a machine. Accordingly, we are usually confronted with two types of interfaces in DSEs: the Graphical User Interface (GUI), and the Application Programming Interface (API). GUIs are a central means of communication between human users and machines. They are central to research as many researchers are not accessing the data produced by a digital scholarly edition directly, but prefer a graphical layer that presents research data for reading, studying and analyzing. The GUI lets the user navigate through the research material and the web presentation built around it. However, it has to be remembered that an interface (GUI or API) of a DSE is always closely linked to the data model of the underlying data and the editorial principles expressed in this data model, in that regard it is a form of pre-selective data management. Interfaces are an interpretation of knowledge and provide users with a more or less “guided tour” through the data and its general presentational setting. Furthermore, they allow the user to answer research questions and aim at supporting the generation of knowledge. Over the last decade the already mentioned APIs have gained importance for DSEs as editors increasingly see their editions not only as static texts published online, but as data that can be linked to other data to answer interesting research questions. The API allows data exchange on a machine to machine level which results in a “guided tour” from DSEs that can be aggregated, interlinked with each other and used to address further research questions by other agents.

As early as the turn of the century, Bethany Nowviskie aligned the highly technically connoted term “interface” to digital scholarly editing, demanding in the middle of the initial enthusiasm a stronger theoretical reflection of the possibilities the (then) new media posed. A critical examination of the topic has also been demanded very early by Jerome McGann in 2001 (171): He witnessed the slow change from a bibliographical to an interface culture. McGann argues for a meaningful exploitation of the familiar media and an aesthetic digital conversion process since, according to him, scholarly book and digital culture do have much in common. What he is missing, however, is the reflective capacity of the digital tools and it is here that he would anchor the potential of the interface. Years later, the critique that the user interface is broadly neglected in the conception of the DSE was verbalized by Hans Walter Gabler in 2010 (48). For him, the user of an edition is still trapped in the traditional receiving role. Editors do not see the user as an equal or peer, and participation in or interaction with a digital edition is not a task available for the user. Envisioning the DSE as the future medium of scholarly editing, Gabler considers the active user involvement to be highly significant. The challenge of social editions, however, lies not in the technical difficulties, as interfaces and workflows are already available, but revolves around theoretical and methodological questions (Brumfield; Robinson, “Theory of Digital Editions” 122).

Similarly, Michael Sperberg-McQueen (30) sees editors migrating from an unruly but well-known (print) to a chaotic, unpredictable environment (digital). One of his solutions to get back a solid user base is to provide problem-solving interfaces. In order to meet these needs of humanities scholars Roberto Rosselli Del Turco (“After the Editing”) was one of the first to introduce a set of design elements DSEs should provide beyond general interface functionalities: these include for example hypertext functionalities, special character handling, image manipulation, advanced search, and complementary data manipulation tools. The demand for a theoretical approach towards the roles the interfaces of a digital edition have to offer was taken up more recently by Patrick Sahle (“Scholarly Digital Edition” 159f.) who points out that as a digital presentation an edition is no longer just data but also design and program code, and in a digital edition one cannot live without the other. A theory of digital editing would have to assess the importance, define the relationship, and estimate the interdependency of data and interface(s), i.e. content and form. It would also have to consider the interaction between historical source, editor and user. Sahle, like others before him, strongly calls for systematic research into this field and expresses the need for a steady development from practise towards theory building. In this context Elena Pierazzo (186-192) points out the importance of digital preservation of DSEs: archival storage of an edition’s data is technically no problem at all, but in relation to interfaces this statement proves much more problematic as it not only involves data standards but also diverse and potentially conflicting versions of different software or even hardware architecture. Since interface designers, and not the editors themselves usually make the interface, they emphasize aesthetic aspects. Pierazzo also investigates the role of aesthetics in GUI design for usability of DSE and concludes that both stability of data and a GUI designed with a certain uniformity are the most cherished factors for a user-oriented presentation. Insecurity in embracing the new possibilities often results in a poor interface and in an annoying user experience (Rosselli del Turco, “Battle” 230). To free the user from the passive consumption of the GUI Johanna Drucker suggests that the interface should not be seen as an object. She promotes instead a sustained, interpretative engagement with the data, the purpose of which is to inspire thinking and generation of knowledge. Therefore, she suggests:

…multiple points of view, correlatable displays, aggregated data, social mediation and networking as a feature of scholarly work, and the qualities of games with emerging rule sets. (§35)

For her an

Interface is a space of affordances and possibilities structured into organization for use. An interface is a set of conditions, structured relations, that allow certain behaviors, actions, readings, events to occur. (§31)

A solution that would free the editor of the burden to provide a GUI, according to Sahle (Digitale Editionsformen 37), could be the provision of the editorial output as mere data via APIs - it would be up to the user to access and analyze this data by her own means, or third party organizations to offer structured forms of access. In the context of interfaces, DH researchers nowadays discuss working with the data at hand, visualizing text and meta information, analyzing and exchanging data, and, of course, the edition as a socially collaborative effort (e.g. Siemens et al.) and also a research commodity (Robinson, “Collaborative Digital Editions”).

The conclusion from this historical overview seems to be that the book paradigm, which was strongly discussed in the early treatises, and the book itself which was often stylized as a feared opponent to early DSEs are no longer the main concern of the digital editor. The digital edition seems to have fully embraced the possibilities of the new media. Even if this is no longer so strongly problemized in scholarly discourse, in the assessment of Joris van Zundert (103-106) the problem still persists and the majority of contemporary DSEs are mere metaphors of books a long way from even utilizing the possibilities of hypertext. He relies on Peter Robinson’s (“Theory of Digital Editions” 123) distinction of editorial approaches “text-as-document” vs. “text-as-work” to contrast recent developments. Like Nowviskie or McGann at the beginning of the century, and so many others in the years to come Zundert again strongly calls for an intensified methodological discussion.

The contributions in this volume build on these conflicting perspectives. Experts of DSEs and Interface Design, editors and users of editions, web designers and developers discuss the relationship between digital scholarly editing and interfaces. In this context the conference team provided a broad selection of topics as intellectual incentive. The discussion was meant to include the critical reflection of the (graphical/user) interfaces of DSEs as much as conceptualizing the digital edition itself as an interface.

How can DSEs take full advantage of their digital environment without losing the traditional affordances that make an edition “scholarly”? What is the role of skeuomorphic tropes and metaphors like footnotes, page turn and index in the design of DSEs and concerning the user interaction?

Do interfaces of DSEs succeed in transferring the complexity of the underlying data models?

Plurality in representation is a core feature of DSE. How do interfaces realize this plurality? Do we need different interfaces for different target audiences (i.e. scholars, digital humanists, students, public)?

How can user interfaces of DSEs succeed in transmitting Human Computer Interaction design principles like “aesthetics”, “trust”, and “satisfaction”?

Citability and reliability are core requirements of scholarly work. Which user interface elements support them? How can we encourage the user to critically engage with the DSE?

What are the users of a DSE actually doing: are they reading the text or searching and analyzing the data?

Can we conceptualize machines as users? How can we include Application Programming Interfaces (APIs) in the discussion on DSEs as interfaces?

Does the development of user interfaces for DSEs keep up with the rising distribution of small handheld devices? Will interfaces on tablets greatly differ from those on computer screens and perhaps encourage a larger readership?

These and other questions on the topic “Digital Scholarly Editions as Interfaces” were vividly discussed during the conference. This volume aggregates twelve of the presented papers that contribute to the debate above and provides a wide range of case studies that highlight the current state of interfaces in DSE. The edited volume is structured according to three methodological perspectives or approaches towards interfaces in textual scholarship: theory, practise and empirical (user) studies. The volume starts with the section ‘Theorizing the Interface’ that includes papers discussing various aspects on the role, importance and needs of interfaces from a theoretical point of view, it continues with the section “The Interface in Practise” containing papers reporting on practical work on the interface of editions. The volume concludes with the section “Evaluating the Interface” in which two papers describe the concrete results and insights derived from user studies.

The first section “Theorizing the Interface” is opened with a contribution by Tara L. Andrews and Joris van Zundert. In literature about DSE the interface is often seen as being secondary and the data is moved to the centre of attention and sometimes this even results in quite hostile remarks about the interface. Andrews and Zundert use this hostility and critique as a starting point to discuss the relevance of GUIs and the key function they play communicating editorial information correctly. Looking at a number of case studies, they carefully assess the arguments and statements editors make about the GUIs of their editions. By carefully using interface elements editors can support an argument, and a clumsy interface design and careless use of interface elements can have negative consequences for the argumentative point an editor wants to make. Therefore, a well-argued and well-arguing GUI is central to the communication between an editor and the user of a DSE.

Wout Dillen continues this line of thinking and elaborates on how the editors can, and should, make a statement using an edition’s GUI. He suggests that the GUI may be seen as the new paratext of a digital edition. Using the Beckett Digital Manuscript Project (BDMP) as a case study, he shows what impact the GUI can have on how an edited text is read. A central point in Dillen’s argument is that depending on what the user is looking for in an edition and to what extend she is immersed in the texts and data, different levels of guidance are needed. Using Dante’s Virgil as a metaphorical guide through a maze, he describes how an editor assists a user to find her way through the maze of digital texts and data provided by a DSE. Without a GUI a user is lost in texts and data, as Dante would have been in the Inferno without the assistance of Virgil.

Shane A. McGarry discusses the issue that digital scholarly editions still follow too closely the printed book in design and functionality. He emphasises the need to consider the interactivity of the interface, which is an advantage over the printed book. In his analysis he focuses on a number of components of digital editions that have been developed originally in a book context and are still used in digital editions. In contrast to the linear reading of the book, the digital medium supports a number of alternative reading modalities. Exploring key literature about digital texts and digital reading, interaction design and information architecture, McGarry suggests that the role of the interface is not only to lead the user to the information she is looking for, but also to engage the reader, to retain and help to recall information. He highlights that in contrast to some arguments that suggest data is important and the interface is not, it has to be kept in mind that attraction is an essential component in a reader’s consideration to use an edition and, hence, of its success.

Ginestra Ferraro and Anna-Maria Sichani take a closer look at project management and design processes used for the development of GUIs for digital scholarly editions. The authors have observed that few DSE projects use project management and product development strategies that are used in the commercial world. This is considered an issue as they could be beneficial for the development of DSEs too. Ferraro and Sichani suggest the integration of design, both as a conceptual framework and as a methodological awareness, early in the development process, in order to assure a better quality product. They outline different software design principles, highlight the relevance of user testing, and discuss an agile-oriented workflow for digital scholarly editing projects. Ferraro and Sichani emphazise that interactions generated by users are an important asset that should be used for future development and can contribute to developing better interfaces for digital editions.

In the final chapter in this section Stefan Dumont discusses the important role of GUIs and APIs in the context of digital editions of correspondence. The editing of correspondence has benefited greatly from digital methods in the past fifteen years. On the one hand, the GUI provides a flexible means for a user to interact with edited letters, much better than printed editions of letters would ever allow, for example, correspondence networks can be visualized and explored. On the other hand, the API provides access to highly-structured data, that can be used for research, shared and connected to other data on the internet. In that context, using the project correspSearch as an example, Dumont emphasises the important role metadata standards play when making data from different editions interoperable and accessible via a single platform.

The second section of the book – The Interface in Practice – looks at practical aspects of interface design. It starts with a report by Roberto Rosselli Del Turco and Chiara di Pietro about the user interface design of the new EVT (Edition Visualization Technology) 2.0, a tool that uses a TEI/XML source document as input to produce a digital edition. In the development of the software tool emphasis has been placed on design of an intuitive user interface. The authors give insights into the design process, discuss problems they faced and what decisions they made to increase the usability of EVT. Besides the design process they also discuss technological challenges and how the software stack changed for the new EVT.

Joshua Schäuble and Hans Walter Gabler discuss interface design in the context of a genetic edition of Virginia Woolf’s “A Sketch of the Past”. The chapter focuses on challenges in the encoding of certain features of genetic editions with the TEI and visualization strategies in regard to the interface. The visualization as such must be realized through (sets of) visualization software. From the construction-in-progress of one such set of modules, the essay demonstrates the design and describes the operation of one modular interface, a Diachronic Slider.

The chapter by Elly Bleeker and Aodhán Kelly also focuses on interfaces of genetic editions. Using a digital museum exhibit, Brulez Digital Exhibit (BDE), as a case study they explore how user interfaces of genetic editions may be designed to communicate complex research results to non-expert audiences. Additionally, the chapter discusses what can be learned from the collaboration between the university and the GLAM sector and how this impacted on the development of the user interface and dissemination strategies of the project.

In his chapter Jeffrey C. Witt argues that too many editions still focus on the GUI and are rooted in the “text-as-document” paradigm. Following the idea of the semantic web, texts can be seen as a series of data points and relationships between them and between data points of other texts. He argues that digital editions should move towards a “text-as-network” paradigm and a clear distinction between the data that can be reused by machines and connected to other data and the presentational layer, i.e. the interface(s) for the human user. The first step to this development has to happen in our heads: instead of thinking first how we want to publish and present the edited texts, we have to think of texts as data first. Using the Scholastic Commentaries and Texts Archive (SCTA) as an example Witt shows the potentials of such a paradigmatic shift. His paper illustrates further how an edition inspired by the “text-as-network” paradigm could look like and that this could result in a plurality of interfaces for various research interests.

Hugh A. Cayless discusses the challenges of user interfaces for critical editions of classical texts and uses the example of the Calpurnius Siculus’ Bucolica edition, which was created as a pilot for a series of new born-digital editions of classical texts by the Digital Latin Library (DLL). The chapter highlights that an edition’s data model and structured data play a central role and form the basis for interfaces. He proposes a web interface using a JavaScript library with the name CETEIcean that directly uses the data model of the edited text. It aims at rich reading environments of digital critical editions and increased usability for scholars.

Federico Caria and Brigitte Mathiak open the last section of the book – Evaluating the Interface – and evaluate the results of a survey and user tests with open task scenarios on three digital scholarly editions: Saint Patrick’s Confessio, Walden: A Fluid Text Edition, and the Emily Dickinson Archive. The goal of the survey and user tests was to gain insight into how end users benefit from DSEs in contrast to paper editions and which kinds of interfaces are more successful than others. Another issue the survey uncovered was that in some editions the user has the feeling of getting lost. Therefore, a minimalistic interface that focuses on the main tasks a user wants or might want to execute could be preferable in terms of usability over an overly complex user interface.

The final chapter also focuses on user studies, but in a different context. Elina Leblanc conducted a user survey on the user interface of the digital library Fonte Gaia. In her contribution she presents the survey results and uses them as a starting point to discuss similarities and differences between the user interfaces of digital libraries and digital scholarly editions. She argues that the three roles of people accessing DSEs (the reader, the user, and the co-worker) can also be translated to the digital library context.

The various exciting contributions to this book and the lively discussions at the conference in autumn 2017 convince us that in order to accommodate as many recipients as possible, DSEs require both a carefully designed, user-centred and taskoriented GUI and a well-documented API that provides access to the data in the edition for further research. In regard to GUI design much can be learned from existing processes and strategies from the design and media industry, even though this certainly requires an even closer integration of the various areas of competence (textual scholarship, digital humanities, design) than is currently the case. As Rucker mentioned in his keynote: the interface design of DSEs has to be a collaborative and interdisciplinary task, that brings together knowledge and skills from different domains.

Many of the projects presented here are ongoing research and highlight the urgency of the topic. These proceedings show the great variety that exists in the approach to and study of this topic. It is to be hoped that the discussion will continue towards a humanities inspired line of thinking about the theory of DSE interfaces.

Bibliography

Bleier, Roman, et al. “Tagungsbericht: Digital Scholarly Editions as Interfaces.” H-Soz-Kult, 17 Mar. 2017, www.hsozkult.de/conferencereport/id/tagungsberichte-7060. Accessed 12 Apr. 2018.

—. “Report: Digital Scholarly Editions as Interfaces Symposium.” DiXiT Blog, 28 Mar. 2017. dixit.hypotheses.org/1250. Accessed 12 Apr. 2018.

Brumfield, Ben. “The Collaborative Future of Amateur Editions.” Collaborative Manuscript Transcription, 13 July 2013,manuscripttranscription.blogspot.co.uk/2013/07/the-collaborative-future-of-amateur.html. Accessed 12 Apr. 2018.

Drucker, Johanna. “Performative Materiality and Theoretical Approaches to Interface”. Digital Humanities Quarterly, vol. 7, no. 1, 2013, digitalhumanities.org:8081/dhq/-vol/7/1/000143/000143.html. Accessed 12 Apr. 2018.

Gabler, Hans Walter. “Theorizing the Digital Scholarly Edition”. Literature Compass, vol. 7, 2010, pp. 43–56. www.academia.edu/214152/Theorizing_the_Digital_Scholarly_Edition. Accessed 12 Apr. 2018.

McGann, Jerome. Radiant Textuality: Literature after the World Wide Web. New York, Palgrave Macmillan, 2001.

Nowviskie, Bethany. “Interfacing the Edition, Talk at the Conference ‘Literary Truth and Scientific Method’.” Charlottesville, Univ. of Virginia. jefferson.village.vir-ginia.edu/bpn2f/1866/interface.html. Accessed 12 Apr. 2018.

OED Online, Oxford University Press, March 2018, www.oed.com/view/Entry/97747. Accessed 12 Apr. 2018.

Pierazzo, Elena. Digital Scholarly Editing: Theories, Models and Methods. 2014, hal.univgrenoble-alpes.fr/hal-01182162.

Pierazzo, Elena and Matthew James Driscoll, editors. Digital Scholarly Editing: Theories and Practices. OpenBooks Publishers, 2016.

Porter, Dot. “‘What Is an Edition Anyway?’ My Keynote for the Digital Scholarly Editions as Interfaces Conference, University of Graz.” Dot Porter Digital, 24 Sept. 2016, www.dotporterdigital.org/what-is-an-edition-anyway-my-keynote-for-the-digital-scholarly-editions-as-interfaces-conference-university-of-graz/. Accessed 12 Apr. 2018.

Robinson, Peter. “Some principles for making collaborative scholarly editions in digital form.” Digital Humanities Quarterly, vol. 11, no. 2, 2017. www.digitalhumanities.org/dhq/-vol/11/2/000293/000293.html Accessed 12 Apr. 2018.

Robinson, Peter. “Towards a Theory of Digital Editions”. Variants, vol. 10, 2013, pp. 105–131.

Rosselli Del Turco, Roberto. “After the Editing Is Done. Designing a Graphic User Interface for Digital Editions.” textitDigital Medievalist, vol. 7, 2011. doi: 10.16995/dm.30. Accessed 12 Apr. 2018.

Rosselli del Turco, Roberto: The Battle We Forgot to Fight: Should We Make a Case for Digital Editions? Digital Scholarly Editing: Theories and Practices, edited by Matthew James Driscoll and Elena Pierazzo, OpenBook Publishers, 2016, pp. 219-61.

Sahle, Patrick. Digitale Editionsformen. Zum Umgang mit der Überlieferung unter den Bedingungen des Medienwandels. Teil 2: Befunde, Theorie und Methodik. Schriften des Instituts für Dokumentologie und Editorik, vol. 8, Books on Demand, 2013. kups.ub.uni-koeln.de/5352/. Accessed 4 Nov. 2017.

Sahle, Patrick. What is a Scholarly Digital Edition? Digital Scholarly Editing: Theories and Practices, edited by Matthew James Driscoll and Elena Pierazzo, OpenBook Publishers, 2016, pp. 19–39.

Siemens, Ray et al. Building A Social Edition of the Devonshire Manuscript. Digital Scholarly Editing: Theories and Practices, edited by Matthew James Driscoll and Elena Pierazzo, OpenBook Publishers, 2016, pp. 137-160.

Sperberg-McQueen, C. Michael. “How to Teach your Edition How to Swim.” Literary and Linguistic Computing, vol. 24, no. 1, 2009, pp. 27–52.

Zundert, Joris van. “Barely Beyond the Book?” Digital Scholarly Editing: Theories and Practices, edited by Matthew James Driscoll and Elena Pierazzo, OpenBook Publishers, 2016, pp. 83–106.

Theorizing the Interface

What Are You Trying to Say? The Interface as an Integral Element of Argument

Tara L. Andrews and Joris J. van Zundert

Digital Scholarly Editions as Interfaces, edited by Roman Bleier, Martina Bürgermeister, Helmut W. Klug, Frederike Neuber, Gerlinde Schneider. Schriften des Instituts für Dokumentologie und Editorik 12. Books on Demand, 2018, 3–33.

Abstract

Graphical interfaces to digital scholarly editions are usually regarded as disconnected from the content of the edition, enough so that an argument has developed against the use of interfaces at all. We argue in this paper that the indifference and even hostility to interfaces is caused by a widespread incomprehension of their argumentative utility. In a pair of case studies of published digital editions, we conduct a detailed examination of the argument their interface makes, and compare these interface rhetorics with the stated intentions of the editors, exposing a number of contradictions between ‘word’ and ‘deed’ in the interface designs. We end by advocating for an explicit consideration of the semiotic significance of the elements of a user interface: that editors reflect on what aspect of the argument their interface expresses, and how that is adding, or perhaps subtracting, from the points they wish to make.

1 Introduction

Some of the tricks of the trade involved in meeting these challenges include studying the design of infrastructure, understanding the paradoxes of infrastructure as both transparent and opaque […] (Susan Leigh Star, The Ethnography of Infrastructure, 377)

In a combative paper presented at the Digital Humanities Conference 2013 in Lincoln Nebraska, Peter Robinson posited: “Your interface is everyone else’s enemy” (Desiderata). He asserted that the very thing which is meant to open up a digital text to users can, rather paradoxically and frustratingly, limit its uses. Infrastructure, as Susan Leigh Star notes, is both transparent and opaque – so long as it works as expected, it is effectively treated as invisible (transparent), but as soon as its affordances or functionality cease to match the needs of its users, those users are at a loss for how, or indeed whether, to continue using it (opaque). Interfaces are themselves a form of infrastructure, subject to the same paradoxical properties. The very purpose of an interface to a scholarly digital edition, whether it be a graphical, command-based, or programmatic interface, is to open up textual information to reader-users. Yet most of these interfaces are designed in a way that renders them neither really open nor neutral. Robinson’s primary complaint was that most interfaces of digital scholarly editions are ultimately nothing but façades behind which textual data is hidden. They proclaim ‘behold… a representation of the text’, but they offer users no further means for downloading data, for reading offline, for adding their annotations, or for interacting in any other meaningful way with the text.

We argue here that the reaction against graphical interfaces for scholarly digital editions, exemplified by but not limited to Robinson’s polemics, is caused by a widespread incomprehension of the argumentative utility of interfaces. Up to now, most interface design has been carried out at the level of “unconscious incompetence” (Wikipedia Contributors, Stages) by textual scholars and the technicians they employ – it is being done, but without much explicit conscious understanding of the impact and effect of particular design decisions. Creators of digital scholarly editions regard interfaces primarily as a utilitarian means of representing the edition, and less often tend to consider the interface as a site of interaction between text and user. We have not developed an explicit understanding of how an interface argues, but such an understanding is necessary to reason about its form, function, and telos. Our purpose here is thus to explore the argumentative aspect of the interface as a first stage in the development of a more consciously-argued approach to graphical interfaces for digital scholarly editions. Our approach is one of critical reception: we will not explore here the mechanisms by which scholarly editions are produced, nor comment on the division of labour that typically goes into the creation of their interfaces. Rather, we will engage with those interfaces on their own terms, as published artefacts oriented toward a particular audience, and examine the messages we “read”.

2 The interface as medium

The user interface of digital scholarly editions is often treated as a content-free and ideally interchangeable appendage to that which is actually considered the scholarly effort or work – the examination and preparation of the text and the scholarly justification for how this preparation was carried out. This is related to the conviction that the interface is, or at any rate should be, a self-contained, unambiguous, non-value-laden digital object that simply transmits a visualisation of digital textual data to a user-reader. At most, its effect is regarded as a visual permutation or aesthetic adornment of the underlying content, the textual data; its purpose is usually to present the text and edition in a way that caters to those who wish ‘simply’ to read the text, or a particular version thereof (although, as we will discover, scholarly editors often produce digital editions that seem to argue against reading).

On a theoretical level, Hans Walter Gabler (47–48) has argued that the ‘autocratic strain traditionally ingrained in the editorial enterprise’ is in part to blame for this attitude towards interface work. On a more pragmatic level, Roberto Rosselli Del Turco (editing, SDE) has pointed out some of the more common flaws in digital scholarly edition interface design. The more computationally-minded in the textual scholarship community clamour, as Robinson did, for machine readable access (APIs) to these editions, in order to apply their stylometric, machine learning, or other such techniques (e.g. Piper, Underwood, Kestemont et al.). Librarians, meanwhile, call for standardization of these interfaces and the underlying data in order to promote interoperability (cf. e.g. Besser).

While the Digital Humanities community engages in its skirmishes about user interface, data access, and interoperability, farther afield under the broad interdisciplinary umbrella of human computer interaction (HCI), the creation and evaluation of user interfaces has grown into an academic expertise of its own, strongly informed by disciplines as varied as graphic design, computer engineering, cognitive theory, and the social sciences (Rogers 2). Vivid debates on the importance of theory formation (e.g. Kaptelinin and Nardi), user experience (cf. Whittaker), and field studies on usability (e.g. Andreasen et al.) drive the field forward. It is thus an opportune time for us in textual scholarship to advance our understanding of interfaces based on this growing body of knowledge.

Here it is useful to point to the work of Alexander Galloway, who understands an interface not as some static digital object but as the effect that results from a dynamic process of transformation or mediation (Galloway viii). As a process of mediation, an interface translates data into different states. Interface effects may be neutral, but more likely they are not, because the processes causing them are usually not impartial automata, but (in a digital context) pieces of software and code whose existence, function, and working were motivated and intentional. As such, interface effects are caused by processes that represent the delegated agency of the persons that designed them (cf. Zundert).

Very little explicit awareness of this dynamic understanding of interface, of the effects caused when interfacing takes place, has crossed over to the literature on digital scholarly editions. How does the look and feel, the visual structure of information, affordances of interaction, or even the aesthetics of a given digital scholarly edition shape the experience of using it? Does the interface promote or discourage a particular mode of reading? Does it suggest or encourage a use beyond straightforward reading? This lack of awareness sits oddly with the point that has been made numerous times, beginning with Cerquiglini, that a scholarly edition is an argument about a text. If it is not particularly controversial to acknowledge that the visual appearance of a text or a picture has a marked effect upon how it is received by an audience – a point that is underscored by the design studies referred to above – then appearance is part and parcel of editorial rhetoric. We argue in this paper that, as producers of these editions, textual scholars need a much greater understanding of how their interfaces are an integral component of the argument they will convey through the act of editing their texts. Although we have some tacit knowledge of this as editors and readers, the field can likely do better at seeking out, and perhaps even producing, empirical information about how the interface – the medium – affects the argument. User interfaces are, after all, a language through which arguments are made, even when the makers of these interfaces are not conscious of the language they are using. As such, they reflect the interpretations of the materials they are supposed to represent as well as the culture, the politics, and the motives of their designers.

Figure 1: La Entretenida and Digital Thoreau digital scholarly editions.

Consider, for instance, the rather striking difference between the initial impressions of two digital scholarly editions: La entretenida by Miguel de Cervantes and Digital Thoreau: Thoreau Digitized. Deliberately. The first interface conveys a fairly conventional ‘scholarly’ perspective, feel, or idea about the text (see fig. 1). The reader can very quickly find his or her way to the features normally associated with printed critical editions: versions consulted, editorial guidelines, a presentation of the text accessible per segment of the play, a hyperlinked index of names, places, occupations, and so on. Four versions of the text are available; the version presented by default is not even that of the editor, but rather a transcription of the first edition with spelling normalised to modern practice (fig. 1).

In contrast, the aesthetics of the interface to the Digital Thoreau work in tandem with its subtitle to provide an experience not merely of the text but also of what it signifies to the editor: deliberation and reserve. This notion or suggestion, however, is very superficial – the edition hides the text to some extent under the layers of aesthetic, but once it has been found, the text turns out to be just as ‘densely scholarly’ as La Entretenida in that it provides multiple instantiations of the text, scholarly footnotes, multiple indices, and so on.

We acknowledge that it is walking quite a fine line to interpret what we see in this way – to describe what we believe to be happening, what it seems to mean and to what it pertains. Should we accept the interface as a utilitarian medium needed only to serve the text in a digital environment, and should we pay minimal regard to presentation? Or should we understand interfaces and their mediality as intentional expressions of the editors’ perspective on the digital edition as a concept, and as a set of deliberate choices about the representation of a text?

3 The interface as argument

Our first observation is that a digital edition’s interface is an argument – not just an argument about the text, but also an argument about the ‘attitude’ of the editor, a window into his or her take on methodology and the digital edition itself. It is also a revelation of the technical skills available to the editor. The interface tells us something not only about the methodology but also about the import of the edition. The Digital Thoreau offers this sort of non-textual stylistic communication in abundance – it argues not just through text but also through the creation of a certain mood with colours, layout, graphics. In contrast, the Cervantes edition makes little attempt to communicate a mood or an emotion; it is clear that these editors would argue that the interface is mostly beside the point, a more or less neutral technical means to an end.

As can be inferred from the attitudes cited at the beginning of this article, the development of an interface for a digital scholarly edition is generally treated as a piece of design independent from the interpretative thrust of the actual content. It is thus considered to lie within the domains of engineering, interaction design, and aesthetics, and, perforce, well outside the domain of textual scholarship. Interfaces are considered essential to communicate content to the user, but they are also usually considered neutral and non-interfering – it is usually taken as a desideratum that they be explicitly divorced from the argument. This othering of the interface can easily be seen in the advice that is usually given to creators of digital editions: that for the sake of sustainability of their research data they should take care to separate content and functionality (e.g. Jannidis, Cayless).

Galey puts this othering in a historical perspective of textual scholarship, pointing to the argument made by DeRose and others: essentially, that a text is the same whether it is printed in Garamond or Times Roman. Others, such as McGann, Hayles, and Kirschenbaum, concerned with the material aspects of text and digitality, contend that meaning and form are ‘distinguishable but fundamentally indivisible’ (Galey 110–112).

DeRose’s argument may well have been informed by pragmatism: there are, of course, very good practical reasons to ensure a separation between form and content. These are primarily bound up in the fact that, up to the present, it is much more technologically feasible to archive static data in the form of plain HTML, XML/XSLT, or even relational or RDF-style database contents, than it is to archive the dynamic functionality of a software or web interface to those contents. Consequently, whatever scholarly content is not cleanly separable from the dynamic or interactive display logic of an edition made today is likely to remain unarchived and, thus, to be lost sooner or later.

This useful and pragmatic practice has, over time, developed overtones of a textual ideology that claims that content and meaning are unproblematically separable from form and function (Galey 110–113). The typical consequential advice to take ‘XMLification’ to the core of the textual scholarship practice, and to put interface work in a peripheral realm of design and engineering, has a theoretical flaw at its core, which is the central tenet of this article. Just as there is no clean separation between data and interpretation, there is no clean separation between the scholarly content of an argument and its rhetorical form (Galey 94). We contend, moreover, that visual display and interactive functionality are an integral part of rhetorical form. The interface is thus an integral part of the argument that an edition makes about a text.

Cerquiglini’s idea that an edition is a theory – and thus an argument – about a text is well-known within textual scholarship by now. Shillingsburg elaborates this idea specifically for the digital sphere: he differentiates between the archive, which collects primary materials and provides access to them while attempting to keep the mediating influence of interpretation to a minimum, and the edition, which will generally include an archive but the primary purpose of which is to provide a scholarly critical argument about the meaning of an archive or how it should be read. Even so, these insights never had a great deal of overt influence on how editions are presented in their printed forms – a reader still expects to find a canonical reading text, one or more apparatuses, and notes provided by the editor.

In the case of digital scholarly editions, the situation can be very different. We start from the observation that there is no single canonical form for a digital edition, although this has become a desideratum for some (Rosselli del Turco, Battle; Czmiel) and although the appearance of guidelines (Rosselli del Turco, editing) and review journals such as RIDE (Steinkrüger) for digital editions and particularly for their interfaces will inevitably have the effect of normalizing certain practices. Nevertheless, a graphical interface (or even an API) is an object that is constructed to present data, which is itself constructed with reference to facts, which themselves are constructed objects.

It is worth pressing this point. In her article on graphical display perhaps most well-known for this very observation, Drucker points out that despite the implication in its very name that “data” (L. dō, dare, datum, “give”) is given naturally by the environment of its production, data creation is in fact a process of active capture of select information by the human or the human-designed algorithm that does the work (Gitelman and Jackson 3). This process of forming and becoming that data undergoes in a scientific context points to a careful selection and argumentation that underlies the presentation of data as meaningful, and as pertaining to a certain argument, as Latour and Woolgar claim (e.g. 255–256). This is not to deny the factuality of data in all or even most circumstances, but it becomes clear in this viewpoint that the data one collects, and the facts one presents, whether underlying the data or derived from it, are not only part of a larger argument but themselves also argumentative constructs.

We can put this in a more concrete perspective by considering the famous example of “bridges with politics”. These are the Long Island bridges designed by Robert Moses in the 1960s, that were supposedly built too low to accommodate public transport buses passing beneath them (Joerges 417–418). The original debate centred around whether the alleged inability of buses to pass underneath the bridges was intended to discourage the circulation onto the island of those racial minorities who were disproportionately dependent on public transportation systems to navigate around the city. That debate remains unresolved; the discussants could agree only on the unknowability of Moses’ specific intentions and motivations, and could certainly not agree on the precise degree to which his civil works furthered the alleged goal of keeping Long Island a de facto white-only zone. The anecdote and the debate serve primarily as a reminder that objects can be agents of politics, but also as an example of how facts can be marshalled in different ways to bring about interpretations of these objects (Woolgar and Cooper 443–444; also cf. Latour Missing Masses). For our purposes, it is worth noting that the objects in question – the bridges – also serve as a physical interface of sorts: one of the disputed claims in the debate is that the freeways over which the bridges were built were the sole practical means of access to Long Island.

Computer code functions similarly as an argument, and as such it is not immune from politics. McPherson has opened the way for an argument about software not dissimilar to that of the Long Island bridges: to what extent, she asks, does the all-white context of the development of platforms such as Unix imply the lack in code of representation of or support for aspects of cultures that lie outside the white-majority mainstream? Although her article was intended less as a substantial accusation toward the Unix programming community than as a provocative talking piece, it has served to pull the curtain from an oft-perceived impartiality or neutrality of code and software. It would be naive at best to regard code as neutral on the grounds that it has at its very base a mathematical nature. In addition to that mathematical nature, it also has a rhetorical one. Pretty much since its inception, it has been argued that code has a literacy (e.g. Kay) that allows a programmer to wield a computer language as intentionally and as meaningfully as any other semiotic system. This eradicates any perceived boundaries between the writing of code and conventional authorship. Moreover, the ability of code to exert control may be far greater than that of conventional text, due to its executable nature that, in contrast to inanimate objects, allows it to adapt and react to specific circumstances (Zundert 365–366). Code, more than bridges and literature are, is a form of delegated agency, and there is little reason to assert that programmers suffer from less intent and bias than any other human being.

We can thus see that code is a construction that, at the very least, furthers, or engages with, a particular set of interpretative perceptions. Data is a construct, built in the very process of its generation. Even facts themselves are constructed, marshalled and interpreted to support or undermine argumentative propositions. How then can a user interface to a digital edition not be a constructed thing, with interpretations and intentions built in from the beginning? Even if a particular piece of code, software, or interface is not meant deliberately to exercise control or to effect certain policies, the production of code and interface remains a thing that is situated, that is: it is built in a context and by people endowed with a certain history, convictions, and cultural identity.

Cultural and historical situatedness thus motivate the development and structure of the interfaces we put on (digital) texts, whether overtly or covertly. This makes these interfaces non-neutral artefacts of the scholarly or technical work rather than neutral intermediaries; that non-neutrality is arguably amplified rather than mitigated by aesthetics. When building interfaces, we generally fail to account for these aspects of interface, and as a result we often ignore the argumentative aspect of the user interface that we provide.

4 The interface as language

And yet. Perhaps the greatest innovation of the digital space is that it gives us a tangible means to express our argument and theory about a text in ways that are not only not limited to the textual, but also not limited to the static. We constantly sing the praises of the possibilities granted to scholarship by these new forms of expression in digital space, and so it is important to set out what these comprise – linear, hierarchical, graph, or perhaps even time-lapse models of the text of the edition, which text can be represented as running text, as text alternatives, as variant graphs, or as tables; the inclusion of imagery or sound that may or may not pertain specifically to a particular portion of text, or alternatively to the argument that the editor wishes to make about the text; the ability to present these texts, images, and sounds either in a static way or dynamically, so that they vary in prominence or even in content, according to what the editor wishes to emphasize or what choices of emphasis the editor has allowed the user to make.

All of these choices, all of the decisions not only concerning the textual content but also the entire experience of its context, are determined by, and determine, what argument, theory, hypothesis, or association the editor has chosen to present. Just as there is not a single data format that will be able to satisfy all use requirements (Vitali), it is difficult to imagine that there can be one universal and universally satisfying interface for a scholarly edition, even when a shared underlying model or encoding standard is used.

At this point, we have reached something of an impasse. It is now clear that any interface is inevitably developed, and its arguments will be couched, in a particular semiotic environment that is extensively shaped by the cultural contexts of the author of a text, its editor, and its audience. The elements of user interface – visuals, colour, dynamic interaction – belong to this semiotic environment and, as such, constitute in themselves a kind of language. As yet, however, we only dimly understand how this language of interface works, what its argumentative properties and aspects are, for it has barely begun its development.

This is not specific to textual scholarship or to digital scholarly editions. Graphical user interface and interaction design work at a complex intersection of visual elements, written language, interaction, aesthetics, and the performativity of software. Although much research has been done on the usability and interaction aspects of this complicated mesh of communicative technologies (cf. Soegaard and Friis), very little theory has emerged about how it constitutes argument. Galey and Ruecker have attempted to make inroads into this problem, and Vanhoutte’s ideas on minimal and maximal editions may also pertain to a theory of how interfaces argue in textual scholarship. Vanhoutte makes clear in any case that the argument is not limited to the written language: simplicity or complexity of interface, for example, argue a level of accessibility for a given edition.

What is safe to assume or conjecture, though, is that different arguments and different critical frameworks should be expected to lead to different interfaces. As soon as we, the editors, consider our intentions for an interface to our editions, the user requirements and certainly the aesthetics begin to differ, and even conflict with each other. This conflict is, moreover, perfectly justifiable as the representation of the various possible arguments about our texts. While a particular group of scholars may agree, for example, on a particular mark-up model or a computational object model as a good common representation of their texts, the interface preferences of each will be an expression of what they individually intend to do – what argument they intend to make – with that model.

For instance, the presence – and even more, the prominence – of an interactive collation tool is linked to the argument that the collation is a changeable thing that should be left to the scholar-user to modify and interpret. Should the text be presented in the form of a graph? Such an interface relates to the argument that the text constitutes a sort of network that should be of interest to the reader-user. Yet another form may give pre-eminence to APIs and incorporate Jupyter Notebook and D3 based visualizations through which these APIs may be explored and used to tweak the text at will, which stresses a meta-argument: that the editorial argument rests solidly in the constitution of the data and its means of access. According to these interface choices, either the user is expected to analyse the text rather than simply reading it, or, at the very least, the editor is disclaiming any right to impose a web-based graphical interference in how the user chooses to read the text.

The possible interfaces for a scholarly edition can thus vary, sometimes radically, despite the inherent validity and fitness for purpose of each of them as expressions of the same underlying model. The situatedness of the scholarship that produced the edition is behind that variation: scholarly argument, aesthetics, human-computer interaction, and usability all contribute to complicate matters more than they help to establish some uniform ideal access to the text.

5 How do interfaces argue? Two tentative case studies

By now it is clear that interfaces argue in a culturally-induced form of symbolic language. Yet we hardly understand what constitute the verbs, nouns, and syntax of this language that is textual, visual, and interactive all at once. One possibility to explore how this language works in the case of digital textual scholarship is to observe and analyse it using methods similar to those of HCI, interaction design, and usability studies. We do so with respect to two case studies of digital scholarly editions. These case studies are not by any measure a fair critique of some of the interfaces that have been built for individual digital scholarly editions. We fully acknowledge that the vast majority of scholarly editorial projects must get by on shoestring budgets, or sometimes on no budget whatsoever. It is therefore often the case that editors have specific wishes for the graphical interface of their editions that cannot be accommodated by designers, engineers, and usability experts within their severely strained financial and temporal budgets. Moreover, when it comes to digital scholarly editing, digital scholars are more often than not pioneers in a quickly changing technological landscape. Many editions based on certain digital technologies may be considered the epitome of what is possible when published, only to be regarded as ‘old fashioned’ when it comes to visualization almost overnight. For these reasons alone, it would not be fair to regard our discussion of these use cases as a sincere critique.