A Hydra-like Russian Doll:
Appraising and Describing
the Shared Drive of
a Staatliches Schulamt
introduction
The President of the Federal Archives of Germany, Michael Hollmann, commented
four years ago: 'The idea of massively unstructured data collections with unclear
contexts is a nightmarish one from the archivist's point of view'.1 This nightmare
became reality in the Generallandesarchiv Karlsruhe (a division of the Landesarchiv
Baden-Württemberg) in 2015, when I began work on a solo pilot project to appraise
the shared drive of a Staatliches Schulamt (state schools authority responsible for
overseeing the schools within a particular district). A follow-up project is not
currently planned, for reasons that will become clear during the following analysis.
I came to imagine the fonds as a hydra-like Russian doll, a metaphor that captures its
problematic features: the fonds had many heads (folders on the first level), which
were internally structured like nesting dolls, that is with many subfolders. It must be
noted that many different nesting structures could be found within the one high-
level folder, at different sublevels, and in different subfolders.
The administrative history of the state schools authority can be described as follows:
several years ago two separate schools authorities were combined into a single
authority, by subsuming one into the other. The successor body continued to run
some of the files of the subsumed schools authority in addition to its own. There was
no DMS (Document Management System) and no employee with an overview of
the entire system. As a result, there was no common approach or consistency
between the various employees in digital record-keeping practices. Since the
authority's personnel were often out on work trips, it was important for their
representatives to have rapid access to their files when dealing with telephone calls.
Therefore the authority developed its own system, in which there were folders with
general access, but also folders belonging to individual employees to which the
others had no access.
I presented a shorter account of my project at the AUdS conference at the
Fachhochschule (University of Applied Sciences) in Potsdam in March 2016. This
was before Victoria Sloyan's fascinating overview of the appraisal of two non-shared
hard drives at the Wellcome Library appeared in November 2016.2 The Wellcome
Library also could not have had access to our data for their analysis, since they were
only available in German. It is therefore very interesting to compare these two
completely independent projects. As we shall see, the contrast suggests a hypothesis
about the appraisal of shared versus individual-user drives.
To begin with, the Wellcome Library project should be described: two external hard
drives belonging to two different scientists were used 'as testing grounds' for
appraisal methods, to develop 'procedures that can be applied to all born-digital
records regardless of format or quantity'.3 It is important to emphasize that each
hard drive came from an individual scientist, and not from a government agency
with multiple contributors to the drive. A significant amount of internal
organization was present in each hard drive - one in particular is described as
'very well organized'- though one hard drive was more hierarchically structured than
the other.4
1. Our schools authority fonds' features
Our fonds contained 57000 files and 6700 folders, 18 GB altogether. As a
comparison, the appraisal project at the German Federal Military Archives that
Frank Bischoff described in 2014 involved 4 GB, 4000 files and 78 folders.5 Our
project therefore represented a new order of magnitude for digital appraisal projects
in Germany.
Folders with the same name and/or identical contents appeared several times on
approximately the same horizontal level as well as on different sublevels.
Additionally, the same files could be found - numerous times - in apparently random
folders and subfolders in different parts of the fonds. Files that belonged together
were also located in different, sometimes widely separated folders. This was
particularly alarming with regard to legal matters: for example, an accusation of a
serious crime and its retraction were saved in totally different places, without any
cross-reference to each other. Even files that had no connection with the work of the
schools authority - such as a recipe for smoked salmon rolls - could be found in the
fonds.6
The folder structure was extremely problematic. There were up to fifteen subfolders,
often without a visible reason. The number of levels varied wildly (sometimes there
was only one subfolder). A particular challenge was posed by the zip-compressed
subfolders with, in part, newer and/or more complete content in comparison with
the 'main' folder. This content had to be extracted and compared file for file with
the 'main' content in order to identify the newest version. The content of a folder
(the individual files) corresponded to the title only in rare cases. There were many
obsolete file formats, and the files concerned could, in part, no longer be opened.
Unsurprisingly, there were no Geschaftsgangsvermerke (filing notes which show
what administrative actions have been carried out with the documents) such as one
would expect in a paper registry system. This meant that it was impossible to
reconstruct what had been done with a particular document - or, indeed, whether
it had been used at all or instead remained in the fonds as a draft.
isabel taylor
150
isabel taylor a hydra-like Russian doll: appraising and describing the shared drive
of a staatliches schulamt
1 M. Hollmann, 'Was wollen wir archivieren?' in: P. Klimpel, J. Keiner (eds.), Was bleibt?: Nachhaltigkeit
der Kultur in der digitalen Welt (Berlin 2013) 190.
2 V. Sloyan, 'Born-digital Archives at the Wellcome Library: Appraisal and Sensitivity Review of Two Hard
Drives', Archives and Records 37:1 (2016) 20-36.
3 Ibidem, 21.
4 Ibidem, 26.
5 F. M. Bischoff, 'Bewertung elektronischer Unterlagen und die Auswirkungen archivarischer Eingriffe auf
die Typologie zukünftiger Quellen', Archivar 67:1 (2014) 49.
6 Many have asked about the fate of the recipe. It was deleted.
151
