Italian Trulli

Scalable Exploration. Prototype Study For The Visualization Of An Author’s Library On The Example Of 'Theodor Fontane’s Library'.

1. State of research and problem outline

In the course of the continuing upswing in writing process research, source types that document processes of authorial self-organization, notation, or reading have come into focus of literary and cultural studies. These include, among others, notebooks (Hoffmann 2008; Efimova 2018), card indexes (Gfrereis and Strittmatter 2013; Krajewski 2011; Schmidt 2016) and author libraries. While digital presentation techniques based on transcriptions and TEI editions are being developed for notebooks and card indexes (cf. Radecke 2018, Schmidt 2016), the digital presentation of author libraries is currently limited to the provision of either digital catalogs that make library metadata available (cf. Paul Celan’s library) or of simple digital copies, which are offered in a viewer and/or as a PDF download (cf. the Grimm library). These forms of presentation, however, do not succeed in grasping the autographical character of these book collections, whose particularity lies in the reading traces that the author (or heirs, estate administrators, owning institutions, etc.) has left in them. At the same time, the aim of an online presentation of such libraries cannot be an edited text; rather, they are intended to enable the reproduction of creative reading and usage practices.

Three objectives can be phrased for such a presentation, with reference to research needs: 1) the provision of the entire volumes of an author's library; 2) the implementation of (re-)search opportunities; 3) the provision of novel applications with which the profile of the collection and reading patterns emerging in it can be discovered, recognised, and researched.

For objective 1), common techniques for presenting digitized books (mostly based on METS/MODS) are already available; for objective 2), similarly, solutions are being offered (librarian or archival discovery tools). For objective 3), which constitutes the focus of our project, no solutions or conceptualizations have been presented so far, although there have been calls for the supplementation of search-based interfaces by concepts that enable serendipitous encounters (Dörk et al. 2011; Thudt et al. 2012; Whitelaw 2015). Even if the establishment of so-called explore modes has already partly contributed to the fulfilment of these goals, there is still a lack of flexible navigation along the relations between the objects in digital collections (Kreiseler et al. 2017).

2. Project, procedure and corpus

We present a draft of an explorable and scalable representation of an author's library, developed in a research-based rapid-prototyping process. Our prototype study combines the presentation of author’s libraries, which is usually addressed as a philological problem, with design-oriented research on the visualization of cultural collections (Glinka et al. 2017; Dörk et al. 2017; Windhager et al. 2018). With this research we present a prototypically implemented approach of how to digitally represent author’s libraries ( video demo ). The project takes up two impulses: On the one hand, it is committed to distant reading (Moretti 2013), which looks for possibilities of pattern recognition in (parts of) an author’s library; on the other hand the representations should be scalable (Weitin 2017), which raises questions about transitions between different granularities.

The basis for the prototype is the library of Theodor Fontane (1819–1898), which comprises approximately 170 volumes. The significance of this author's library is primarily due to its provenance. It is irreplaceable because of the marginal notes and glosses written inside the books by Fontane and valuable because of numerous presentation copies (Rasch 2005).

In addition to the complete digitization of the holdings according to archival standards, and the indexing of individual volumes according to bibliographic standards, data is collected at both page and corpus level. Access to the data is possible at four levels: overall corpus, object, page, and individual phenomenon (note, comment, etc.). The visualization enables various access and analysis options, which ideally reveal the broad spectrum of thematic and personal clusters and the different uses of commentary within the author’s library.

3. Visualization concept

The visualization places a special focus on continuous navigation through zoom and filters on several granularity levels. This allows for the exploration of individual objects as well as their comparison and broad overviews of the whole inventory. The interactive prototype offers three basic levels for navigating the collection: the level of the individual authors, books, and pages.

Fig.1: Landing page/book level with overview of all books

The starting point for the exploration of the visualization is the book level, which provides an overview of all books in the library, sorted by authors—starting with the author with the most books (Fig.1). Each book is represented by a vertical bar, in which each page of the respective book is represented by a segment, ordered from top (first page of the book) to bottom (last page). The page segments are color coded according to the category of occurring reading traces. While pages without traces are displayed in white, pages containing traces are colored, divided into four categories: 1) provenance (grey tones), 2) markings (red tones), 3) marginalia (blue tones), and 4) additional material (yellow). Hovering over a segment displays a preview of the respective page scan and the book title (Fig.2).

Fig.2: Page level with a book selected, filtered by marginalia, and hovering over an element

Giving an overview and resembling individual bar codes for each book, the book level visualizes the number of pages of a book (total length of the bar) as well as the distribution and patterns of reading traces (color coding). Above the visualization, a filter navigation serves as a legend for the color coding and offers the possibility to focus on certain categories of reading traces. The selection of a filter triggers the unfolding of the corresponding subcategories in the filter bar (Fig.2,3). In addition, a search bar enables specific keyword search in all transcribed marginalia, highlighting respective occurrences in the visualization. By selecting a book, the other books are compressed and a detailed view of the selected book is unfolded, revealing additional metadata and details.

The browser's scroll function can be used to reach the other two granularity levels of the visualization. Following the principles of semantic zoom (Perlin and Fox 1993), scrolling upwards leads to a higher level of abstraction and downwards to a higher level of detail—thus allowing a change between three main granularities: authors, books, and pages. Scrolling enables continuous, meaningful transitions between the views and offers the possibility to go back and forth at one's own speed with the aim of making transitions between views more comprehensible.

Fig.3: Author level, filtered by marks, the colors of various mark types are unfolded

In contrast to the book level, on the author level (Fig.3) all books of an author are combined by presenting the total distribution of the reading traces in form of an area chart. It is therefore possible to compare Fontane's reading traces, distributed over the works of different authors. However, the higher abstraction also allows insight into broader patterns.

Navigating from the book level towards the other direction to the lower page level (Fig.2), the page fragments of books of a selected author are zoomed in, enabling interaction with individual reading traces. Furthermore, selecting a book on this level unfolds a visualization of the transcribed marginalia.

All selections, filter options and the selected level of granularity are encoded via a fragment identifier in the URL, allowing the sharing and bookmarking of specifics views and the use of the browser's history functions.

4. Reflection of results

The novel visual approach to author’s libraries developed within the framework of the prototype combines design-oriented approaches to the visualization of cultural collections with philological, archival, and library research questions. By means of visual filters, subsets can be identified and categories can be created that enable pattern recognition within the collection. Users should be enabled to grasp concepts and themes within dedicated categories and across their boundaries. The focus of the project lies on the integration of search functions, visibility and scalability in visualizations, and how further exploration possibilities can be opened up by different access options to a collection.

The development of new research questions during the prototyping process and the associated readjustment in the indexing process illuminate the interactions between visual research, metadata management, and philology. It became clear that the collection is to be understood as a construct that is only generated retrospectively. In the process, 'the' collection emerges from the unity of perspective and object. The resulting consequences of offering multiple views of 'the' object(s) are reflected in the various granularity levels of the prototype and result in an observation-dependent visualization of the available source data.

Appendix A

Bibliography
  1. Dörk, M., Carpendale, S. and Williamson, C. (2011). The information flaneur: A fresh look at information seeking. Proceedings of the SIGCHI Conference on Human Factors in Computing Systems. ACM, pp. 1215–1224.
  2. Dörk, M., Pietsch, C. and Credico, G. (2017). One view is not enough. Information Design Journal, 23(1): 39–47.
  3. Efimova, S. (2018). Das Schriftsteller-Notizbuch Als Denkmedium in Der Russischen Und Deutschen Literatur. Vol. 22. Paderborn: Wilhelm Fink.
  4. Gfrereis, H. and Strittmatter, E. (eds.) (2013). Zettelkästen. Maschinen der Phantasie. Ausstellungskatalog. Deutsche Schillergesellschaft: Marbach a.N.
  5. Glinka, K., Pietsch, C. and Dörk, M. (2017). Past Visions and Reconciling Views: Visualizing Time, Texture and Themes in Cultural Collections. DHQ: Digital Humanities Quarterly, 11(2).
  6. Haber, P. (2010). Autorenbibliotheken im digitalen Zeitalter. In Quatro. Zeitschrift des Schweizerischen Literaturarchivs 30/31: 39-43.
  7. Höppner, S., Jessen, C. and Münkner, J. (eds.) (2018) . Autorschaft und Bibliothek. Sammlungsstrategien und Schreibverfahren. Mit einem Vorwort von Reinhard Laube. Göttingen: Wallstein.
  8. Hoffmann, C. (2008). Wie lesen? Das Notizbuch als Bühne der Forschung In Griesecke, B. (ed.): Werkstätten des Möglichen 1930-1936: L. Fleck, E. Husserl, R. Musil, L. Wittgenstein, Würzburg: Königshausen & Neumann, pp. 45-57.
  9. Knoche, M. (ed.) (2015). Autorenbibliotheken. Erschließung, Rekonstruktion, Wissensordnung. Wiesbaden: Harrassowitz.
  10. Krajewski, M. (2011). Paper Machines. About cards & catalogs, 1548-1929. Cambridge: MIT Press.
  11. Kreiseler, S., Brüggemann, V. and Dörk, M. (2017). Tracing exploratory modes in digital collections of museum web sites using reverse information architecture. First Monday, 22(4).
  12. Moretti, F. (2013). Distant Reading. London: Verso.
  13. Perlin, K. and Fox, D. (1993). Pad: an alternative approach to the computer interface. Proceedings of the 20th Annual Conference on Computer Graphics and Interactive Techniques. ACM, pp. 57–64.
  14. Rasch, W. (2005). Zeitungstiger, Bücherfresser. Die Bibliothek Theodor Fontanes als Fragment und Aufgabe betrachtet. In Schneider, U. (ed.): Imprimatur. Ein Jahrbuch für Bücherfreunde. N.F. [Bd.] XIX. Wiesbaden: Harrassowitz, pp. 103-144.
  15. Rohmann, I. (2015). Aspekte der Erschließung und Rekonstruktion nachgelassener Privatbibliotheken am Beispiel der Büchersammlungen Herders, Wielands, Schillers und Goethes. In Knoche, M. (ed.). Autorenbibliotheken. Erschließung, Rekonstruktion, Wissensordnung. Wiesbaden: Harrassowitz, pp. 17-59.
  16. Schmidt, J. (2016). Niklas Luhmann´s Card Index: Thinking Tool, Communication Partner, Publication Machine. In Alberto Cevolini (ed.). Forgetting Machines. Knowledge Management Evolution in Early Modern Europe. Leiden: Brill, pp. 289-311.
  17. Thudt, A., Hinrichs, U. and Carpendale, S. (2012). The bohemian bookshelf: supporting serendipitous book discoveries through information visualization. Proceedings of the SIGCHI Conference on Human Factors in Computing Systems. ACM, pp. 1461–1470.
  18. Wieland, M. (2015). Materialität des Lesens. Zur Topographie von Annotationsspuren in Autorenbibliotheken. In Knoche, M. (ed.).  Autorenbibliotheken. Erschließung, Rekonstruktion, Wissensordnung. Wiesbaden: Harrassowitz, pp. 147-173.
  19. Windhager, F., Federico, P., Schreder, G., Glinka, K., Dörk, M., Miksch, S. and Mayr, E. (2018). Visualization of Cultural Heritage Collection Data: State of the Art and Future Challenges. IEEE Transactions on Visualization and Computer Graphics.
  20. Whitelaw, M. (2015). Generous Interfaces for Digital Cultural Collections. DHQ: Digital Humanities Quarterly, 9(1).
Anna Busch (annabusch@uni-potsdam.de), Theodor-Fontane-Archiv / Universität Potsdam, Germany und Mark-Jan Bludau (mark-jan.bludau@fh-potsdam.de), Urban Complexity Lab, Fachhochschule Potsdam, Germany und Viktoria Brüggemann (viktoria.brueggemann@fh-potsdam.de), Urban Complexity Lab, Fachhochschule Potsdam, Germany und Kristina Genzel (kristina.genzel@uni-potsdam.de), Theodor-Fontane-Archiv / Universität Potsdam, Germany und Sabine Seifert (sabiseif@uni-potsdam.de), Theodor-Fontane-Archiv / Universität Potsdam, Germany und Peer Trilcke (trilcke@uni-potsdam.de), Theodor-Fontane-Archiv / Universität Potsdam, Germany