the wildly varying degrees of importance of the documents within the fonds that
I gained during the later appraisal reinforces this decision in retrospect.
I also decided not to ingest the data into our digital repository and appraise after the
fact, based on 'descriptive metadata generated through cataloguing' as suggested by
the Paradigm Project, an idea that was also rejected by the Wellcome Library.39 As
Sloyan notes, it seems unwise 'to ingest files into [a digital storage system] that are
then going to fail appraisal, as they then need to either be deleted or left in the
system for no real reasons' - deleting files from the system and their corresponding
entries in the catalogue database, or leaving the files there, can both lead to wastage
of resources.40
I therefore found myself in precisely the situation that Peter Sandner predicted at the
AUdS conference in Potsdam in 2007: because of the impossibility of achieving a
sufficient overview of the fonds, I was unable to avoid individual file autopsy.41
Vera Zahnhausen from the Federal Archives of Germany has also noted that when
folders are stacked one within another to a very deep level, it is not possible to
recognize which level corresponds to the administrative procedure and which
represents the 'file' (Akte) level: one is therefore forced to 'open and view all the
folders and all the individual files on every level'.42
The Paradigm Project, however, contends that this approach is impossible: '[in a case
such as ours] lack of structure leaves item-level appraisal as the sole option and the
volume prohibits this'.43 I therefore decided to conduct a 'time and motion'
study to test this thesis. The only way to determine the time needed for such a project
is to actually appraise the fonds: only then does the chaos that lurks beneath become
obvious.
5. The final approach
I continued to use our appraisal model for schools authorities, although it is
oriented towards well-structured paper records. It formed a point of departure,
although, as already noted, the functions in this case were to a significant extent out
of date and new functions could be recognised, sometimes from the fonds itself and
sometimes via discussion with the authority's staff. The new method involved
orienting individual files, rather than folders, to functions - in other words, I
appraised according to content that related to these functions. Where there were
clearly new areas, their archive-worthiness was determined in consultation with my
supervisor Dr. Jürgen Treffeisen. This resulted in a long list of more than twenty
pages containing new functions and our decisions: accession or destroy.
5.1. The concrete workflow
1. Create a backup copy.
2. Use Fast Duplicate File Finder to de-duplicate. I always kept the newest
version, because there was a higher chance that this was the version that the
authority actually used for its work (this corresponds to Sloyan's approach).44
The result was 7559 duplicate file groups and 18297 duplicate files, which
meant that 32% of the entire fonds consisted of duplicates. Fast Duplicate
File Finder is excellent, but it does not always choose the oldest version for
deletion, so it was necessary to check the entire list of results and move the
check marks where necessary. In the overwhelming number of cases it was
impossible to discern a rationale for keeping duplicates of the files, unlike in
Sloyan's project, in which it was clear that certain duplicates were, for
example, supporting materials for meetings.45 Since I came to the conclusion
that in the vast majority of cases it was impossible to identify any role played
by the file in the new context, I decided to consistently delete all duplicates.
3. Then, because deduping to this extent is bound to result in empty folders, I
used Remove Empty Directory. 279 empty folders were deleted.
4. After these preparations, I appraised on the file level. Our version of
Windows (7) has a useful preview function in Explorer, so that I was able to
scroll through the files, quickly examine each, then delete it if necessary
(my speed-reading ability was of considerable help in this situation, but this
may also mean that the end result, in terms of time needed, may not be
universally valid). Files in obsolete formats were deleted on the ground that
they were not archive-capable (archivfahig).
5. When I had appraised all the files within a folder, I wrote 'appraised' in the
folder name, so as to avoid redundant appraisal: a real risk when one is faced
with thousands of folders and multiple sublevels of folders.
6. I was often faced with the difficulty that a file was present as a .pdf and .doc,
and in that case I simply kept the newest version, even when it was in .doc
format.
7. After appraisal I had to use a tool to remove all the appraisal notes from the
folder names. For this I chose Bulk Rename Utility. It is extremely fast and has
numerous, highly effective search and replace options, so that it is well worth
investing the time in getting to know the interface. It was still necessary to do
some manual renaming afterwards with the search function in Windows, but
renaming 2032 folders only took 8 hours in total.
The result of the appraisal was that I kept substantially less data: 10884 files opposed
to, originally, 57000; 2032 folders as against 6700; and 5.6 GB instead of 18 GB.
This means that I deleted almost 70% of the fonds. This may seem a high proportion,
but it represents a generous retention quota compared with comparable appraisal
endeavours in the analogue world.
5.2. The description
For the description, my colleague Dr. Kai Naumann of the Staatsarchiv Ludwigsburg
suggested reflecting the folder structure by using subheadings in the Scope
archival database, but only down to sublevel and subfolder/subfile 4. The folder, or
alternatively the file, on sublevel 4 was then designated the unit of description.
The end result was 359 individual entities. The numbers of subfolders and files
within each entity were provided in the item description. Thus, for example, the
archival entity 'Intern', which in the folder and Scope subheading structure comes
praktijk
39 Paradigm Project; Sloyan, 24.
40 Sloyan, 24-25.
41 P. Sandner, Bewertung digitaler Aufzeichnungen aus dem Dokumentenmanagementsystem: Gratwanderung zwi-
schen willkommener Automatisierung und langwieriger Einzelbewertung (2007) 8, http://www.staatsarchiv.
sg.ch/home/auds/11/_jcr_content/Par/downloadlist/DownloadListPar/download.ocFile/Text%20
Sandner.pdf.
156
isabel taylor a hydra-like Russian doll: appraising and describing the shared drive
of a staatliches schulamt
42 V. Zahnhausen, 'Überlieferungsbildung von analog zu digital: Erfahrungen bei der Übernahme von digitalem
Archivgut', in: K. Riemann (ed.), Bewertung und Übernahme elektronischer Unterlagen - Business as usual?:
Beitrage des Expertenworkshops in Münster am 11. und 12. Juni 2013 (Münster 2013) 1.
43-Paradigm Project.
44 Sloyan, 26.
45 Sloyan, 26.
157