Interpretation of Digital Records

Collaborative notes taken by paricipants of the DH Summer School Switzerland:

Download as PDF

Tutorial/Workshop Session 3: Guido Koller (@GuidoKoller) (Swiss Federal Archives): Interpretation of Digital Records: The Swiss Federal Archives’ case

According to Koller, problems relating to archiving/preservation and metadata management are at least partly solved. Still at stake are however questions on how to make records best available and how to interpret digital records.
Archives have a 30 year ceiling, newer documents are available for consultation on request
Looking at the archive, how it works with digital records
He discusses search problem with OCR (e.g. numeric data/statistics that are printed as text strings)
short intermezzo: DODIS search (
Archives in the digital age (BAR perspective)
1980s: digital born data created in the administration
– oldest data: statistical office; magnetic tapes are lost (hardware not available, magnetic charge lost)
1990s: digital finding aids
2000s: digitalization, archiving digital data
– OAIS standard
2010s: big issue: context information
– until now essential distinction between primary data and metadata
– fulltext searchching makes metadata less important; for contextualisation however metadata matters
– (relational) databases are flattened (SIARD tool)
User needs
– data driven analysis; PDF is not very suitable for data analysis
– worst to best quality: Pdf->Excel->CSV (Comma-separated value)->XML- >Permanent Link/linked data
– collaborative web (e.g. user generated comments)
– developing infrastructure for access
– organizing spaces of information (“proto VRE”)
– building up knowledge (internally to meet the challenges, externally as a data contributor)
– metadata description on file or document level? consensus: the level is less important than the quality
– federated search on several archives/single point of access desirable
– abandon absurd rules (3 pictures of manuscript material per day and user in one cantonal archive)