Amalia : an eSciDoc based solution to manage the production, processing and publishing workflows of TEI data.
Résumé
At the ENS Lyon, various teams have chosen TEI to represent their data - corpora, critical editions, etc., with the benefits we know in terms of expertise sharing, strength of XML technologies, long term preservation ... However, the use of the standard is not sufficient to avoid a certain complexity when projects are growing : with the issue of file versioning (which can of course always be solved by an ad hoc software) appear issues of dependence between these versions : for instance, when a XSL file has been modified, which data files have to be reprocessed? What are all the other operations to be launched ? This complexity is increased by the problems of authentication and authorization, and by the great variety of software tools involved in the production, exploitation, visualization and / or publication of these data. As part of a sharing policy within the ENS, we develop 'amalia, a solution based on the platform eSciDoc to address this problem of complexity. eSciDoc is a data management system which provides functionalities to organise data, associate metadata to data, manage versioning, and control access. 'amalia is a web application which elaborates on eSciDoc and allows the definition of data workflows, i.e. chains of transformations and/or actions operating on data. The actual realisation of the actions orchestrated by 'amalia is not constraint and may rely on web service, local script, ... 'amalia can be conceived as a glue between various subsystems where the data they work on is ultimately managed by eSciDoc. We will propose a demonstration of 'amalia and a feedback on the challenges and benefits of this solution in our own experience. We will first show the first workflow we defined, which was for the project Hyperdonat. This workflow transform from ODT to XML TEI and from XML TEI to HTML and index the XML TEI in the XML database BaseX. We will also show other workflows at various stages of development, and involving other third-party systems such as Dinah.