Domain Driven Design and Domain Specific Modelling for Digital Textual Scholarship

Over the last years, the digital turn and the world wide web have led historical studies towards an automatic processing of their own data and consequently towards new forms of scholarly editing and publications. In this framework, scholars have adopted digital models, electronic elements and computational features in their work, but these new instruments are generally derived from other disciplines. For example, they exploit optical character recognition from image processing, corpora annotation and natural language processing from computational linguistics, text alignment from bioinformatics, text meaning from knowledge engineering, text presentation from data visualization. However, these latter research areas do not cover entirely the specificity of the fundamental requirements of the scholarly domain (for instance, treebank data models do not provide the adequate abstractions to manage multiple variant readings and multiple text interpretations). To exceed these issues, it is essential to adopt correct design approaches devoted to analyze the problem space of the historical source editing field. This rigorous and formal analysis will shape suitable architectures, design patterns, data abstractions and procedural abstractions for the constitutive features of the digital scholarly editions. Moreover, this modelling process will produce generic, flexible, maintainable and reusable digital models and modular textual scholarly environments. This contribution aims at discussing software engineering approaches, within an object-oriented paradigm, towards the definition of domain specific abstractions (DS-ADTs). In this way, it will be possible to accommodate domain needs by formally defining core "unities of concerns" which actually adhere to both the traditional and the digital editorial domain.