Speaker
Description
The Academy of Sciences and Literature Mainz (AdW) produces a number of critical editions of cultural heritage, which become increasingly multimodal over long funding periods of 12 or more years. Historical dictionaries, for example, turn to include geodata and statistics, image archives gain contextualisations and 3D reconstructions, and letter collections require annotations as well as digitised archive records. In addition, the Linked Open Data paradigm defines common formats to make (and keep) the editions' content available in, and federated search providers specify their additional APIs to support. Last but not least, digital editions should be accessible by default rather than as an exception.
To address such challenges, the AdW has tried to produce a fixed set of edition-related extensions for the content management system TYPO3 for several years now – with limited success. They were spread across multiple versions of the platform, hard to combine, and difficult to maintain as they were designed for one project and then heavily adapted for another without further documentation. In this talk I will outline the process of redesigning and rewriting this software stack.
The Cultural Heritage Framework 2 (CHF) is a toolkit for web apps that enable users to edit and publish cultural heritage data. Instead of abstract user stories, which have proven useful for the development of individual editions, the CHF was rebuilt based on media ecology theory. In this process, the software is seen as an entity in relation with various other human and non-human entities which can be grouped into media ecosystems. The product needs to be designed in a way that allows it to ‘survive’ and be seen as a good-faith actor in various such ecosystems, including (but not limited to) academic editorial teams, frequently changing and often inexperienced maintainers, mobile web browsers, content aggregators, and users with accessibility needs.
Compared to previous attempts, this analysis led to a different feature set including an adaptable and atomic user interface, embedded JSON-LD metadata, semantic main classes, standardised search functionality, an import/export mechanism, integrated user documentation, and extensive developer documentation. Through interlocking and coherent components for specific data types, projects using the software may now add functionality as they grow or change focus over time. All components feature interfaces to edit the data by reusing features of the TYPO3 platform, but also allow importing data from other systems that a project may use or exporting the data to TEI XML or triple stores.
The talk focuses on the practical application of media ecology theory in the CHF, which is not yet common in Digital Humanities software. It specifically dives into the consequences of evaluating accessibility software, its users, and entities with more limited requirements as one ecosystem instead of a single user story: accessibility needs to be observed not just on the level of content consumption but also on the levels of data analysis and production.
