La presente comunicación tiene el objetivo de reflexionar sobre el estado actual de las interfaces de visualización de las ediciones digitales académicas y aportar nuevos datos acerca de las posibilidades de publicación de ficheros textuales en el formato XML-TEI. El tema se abordará a partir de la experiencia madurada por el Progetto Mambrino (Universidad de Verona), en el contexto de la creación de una Biblioteca Digital de las novelas caballerescas italianas del Renacimiento, que son traducciones y continuaciones del famoso género español de los libros de caballerías. En concreto, se relatan las distintas fases que dentro del proyecto han llevado a la personalización de la aplicación TEI Publisher para la visualización de las ediciones, teniendo en cuenta también aspectos relacionados con la accesibilidad y la experiencia del usuario.

The Native Bound Unbound project seeks to document the lives of indigenous enslaved people in the Americas. Led by Dr. Estevan Rael-Gálvez, the project aims to tell the stories of the millions of indigenous people who were enslaved since the arrival of Columbus. Our firm, Performant Software Solutions, was selected to develop and implement the data infrastructure and website for Native Bound Unbound. In this presentation, we will discuss our approach, which utilizes technologies based on the TEI Guidelines and the International Image Interoperability Framework (IIIF).

To create the fullest picture we can of the individuals whose lives are being documented, we need to gather information from as many sources as possible. This means we must manage many different types of primary sources, including: baptismal records, census records, court cases, tombstones, and oral histories. Texts are often handwritten in a variety of languages including Spanish, English, and Dutch. From this material, we need to distill information about individuals, events in their lives, and the places they lived and worked.

One possible approach would be to read all this material and simply key it into a database. However, this would not make it possible for other researchers to examine our evidence and draw their own or further conclusions. Furthermore, this project is for the public as much as it is for researchers. So we need to both transcribe and translate the material we are working with so that we can display it on the website alongside the page images. These sources are then linked to the records generated from them and the whole process is open to inspection.

Structuring the texts and marking up the people, places, and events in them allows us to programmatically enter the records into the database. These then need to be scanned for duplication and de-duplicated. Updates to the database or the transcriptions should not cause a loss of data or duplicate data. And, of course, our understanding of what information we want to collect and can collect has been evolving throughout this process, so we need a certain level of flexibility in our data models.

Performant brought to this project an array of open source projects that we have developed for previous digital humanities projects, plus some projects developed by other groups that aligned with our client’s needs. These tools include: FairCopy, Core Data, EditionCrafter, FromThePage, Splink, and our IIIF CMS. These tools, taken together, provide a complete workflow for the data infrastructure of this ambitious project. They also provide an application programming interface on top of which we were able to construct the public facing website. This presentation will provide a detailed look at our technical approach, these tools, and the final project website.

DH research usually works with very diverse datasets and tools: from facsimiles, through HTR or OCR, transcription, collation, entity recognition, authority links to further annotation and, eventually, publication online or in print. So much more is within our easy reach than even just a decade ago and these opportunities are widely embraced in our community. Nevertheless, creation of a digital resource is an intrinsically complex subject, as measured by a random assortment of acronyms from TEI to IIIF, so the average scholar can't be expected to have an overview of the ever fluctuating technological landscape. And yet she is still ultimately responsible for designing and orchestrating complex, collaborative workflows and has to directly face the short- and long-term consequences of each decision.

I will discuss what happened when a community of practitioners instead of accepting the unfavourable situation, worked together firmly prioritizing the development of a generic, systematic solution above the specific, without compromising on the latter, analyzing the connected cases of TEI Simple, TEI Processing Model, TEI Publisher framework, and collectives like e-editiones and Sources Online. I would like to demonstrate the visible impact these collaborations, grounded in the mottos of "power to the editor" and "standardize where you can, customize where you must" had on the landscape of digital scholarly editions in recent years.

With Publisher now in version 9 and an impressive international community of e-editiones, we can make some interesting observations. About 40 projects chose to register on the e-editiones website, similar number is currently in preparation and we estimate perhaps another hundred that we never yet heard about. I will talk about various observable trends among editorial projects that we can trace back to resources - software releases, demos and samples - made available by the community across the years. These range from functional features to adopting the text encoding strategies and data organization practices, to standards and open libraries, e.g. XML vocabularies like TEI, JATS or DocBook; MathML and TEX; IIIF for the images together with OpenSeaDragon and Tify viewers; Verovio for sheet music; OpenAPI for the specification of the programmatic interface and DTS for machine-consumption of digital text collections and so on.

This has a profound effect on establishing a consensual good practice and base level of interoperability for digital humanities projects. For example, each and every TEI Publisher-based application exposes well documented programmatic interface, e.g. DTS protocol or authority registries. When it's available as default and by design integrated into the framework we treat interoperability as a priority and achieve it out of the box. Similarly, the re-use approach led to emergence of specialized data models, successfully adopted wholesale or customized by different projects.

Concluding, the availability of an open source framework which allows users to mix existing solutions, strengthened by a supportive community of practice has proven extremely successful. For a practical demonstration of the power this approach affords, I will discuss the Sources Online project.