Workflow and Information System

The journey from a file in the archive to its digital and HTR-based presentation on an online platform passes through several stations. These steps make up the overall project workflow. They are based on a broad technical infrastructure. The workflow of our project, which is geographically spread over three locations, consists roughly of six main stations:

  1. Preaparation of the files and the process (restorative, archival, digital)
  2. Scanning
  3. Enrichment with structural and metadata
  4. Providing the files for Transkribus
  5. Automatic Handwritten Textrecognition (HTR) with Transkribus
  6. Online presentation in the Digital Library Mecklenburg-Vorpommern

It proved to be helpful that we not only had strictly defined the individual steps in advance but also determined the persons responsible from the beginning, i.a. experts for the individual tasks as well as coordinators for the steps across stations and locations. This ensures that all parties involved always know the respective contact person. Open questions can thus be answered more easily and any problems that may arise can be solved more efficiently.

Especially with the scanning of the Spruchakten, we have not proceeded chronologically from the start. We did not scan the inventory ‘from top to bottom’. Instead, we first selected and edited individual representative volumes between 1580 and 1675. We wanted to create mighty HTR models first. Only then did we ‘fill up the gaps’. This procedure showed us, how crucial it is to continuously document the progress of the project with all its indivdual areas and stages so that it may not become confusing. There are many ways to do this.

We keep – meanwhile very colourful – spreadsheets on the progress of the various collections and files. However, they only depict partial processes and are only accessible to the coordinators. But these spreadsheets have to be maintained and for this purpose the progress in the various areas have to be closely monitored.

Another possiblity is the Goobi workflow. Some tasks take place on the Goobi server anyway. In addition to those we can freely add tasks to the Goobi-workflow which do not have to be related to Goobi itself. All of them can be ‘accepted’ and ‘completed’ on that platform to reflect the progress of the project. However, the condition here is that all project contributors must be familiar with this workflow system. Where this is not the case, an „external“ information system must be selected that everyone can access and handle.

The different partners of our project therefor jointly keep a wiki (e-collaboration).