Day 3 of ECDL started for me with the Query Log Analysis session. I thought perhaps that, now the papers were getting heavily into IR technicalities, I might not understand what was being presented or that it would be less relevant to archives. How wrong can you be! Well, ok, IR metrics are complex, especially for someone new to the field, but when the first presentation was based upon a usability study of the EAD finding aids at the Nationaal Archief (the National Archives of the Netherlands), it wasn’t too difficult to spot the relevance. In fact, it was interesting to see how you notice things when the test data is presented in a foreign language, that you wouldn’t necessarily observe if they were in your mother tongue. In the case of the Nationaal Archief, I was horrified at how many clicks were required to reach an item description. Most archives have this problem with web-based finding aids (unless they merely replicate a traditional format, for instance, a PDF copy of a paper list), but somehow it was so much more obvious when I wasn’t quite sure exactly what was being presented to me at each stage of the results. This is what it must be like to be an archival novice. No wonder they give up.
The second paper of the morning, Determining Time of Queries for Re-ranking Search Results, was also very pertinent to searching in an archival context. It discussed ‘temporal documents’ where either the terminology itself has changed over time or time is highly relevant to the query. This temporal intent may be either implicit or explicit in the query. For example, ‘tsunami + Thailand’ is likely to refer to the 2004 tsunami. These kinds of issues are obviously very important for historians, and for archivists making temporal collections available in a web environment, such as web archives and online archival finding aids.
Later in the morning, I was down to attend the stream on Domain-specific Digital Libraries. One of these specific domains turned out to be archives, with an (appropriately) very philosophical paper presented by Pierre-Edouard Portier about DINAH [in French]. This is “a philological platform for the construction of multi-structured documents”, created to enable the transcription and annotation of the papes of the French philosopher, Jean-Toussaint Desanti, and to facilitate the visualization of the trace of user activities. My tweeting of this paper (limited on account of both the presentation’s intellectual and technical complexity and the fact that I’d got to bed at around 3am that morning!) seemed to catch the attention of both the archival profession and the Linked Data community; it certainly deserves some further coverage in the English-speaking archival professional literature.
In the same session, I was also interested in the visualization techniques presented for time-oriented scientific data by Jürgen Bernard, which reminded me of The Visible Archive research project funded by the National Archives of Australia. The principle – that visual presentations are a useful, possibly preferable, alternative to text-based descriptions of huge series of data – is the same in both cases. Similarly, the PROBADO project has investigated the development of tools to store and retrieve complex, non-textual data and objects, such as 3D CAD drawings and music. There were important implications from all of these papers for the future development of archival finding aids.
In the afternoon, I found myself helping out at the Networked Knowledge Organization Systems/Services (NKOS) workshop. I wasn’t really sure what this entailed, but it turned out to involve things like thesauri construction and semantic mapping between systems, all of which is very relevant to the UK Archives Discovery (UKAD) Network objectives. I was particularly sorry I was unable to make the Friday session of the workshop, which was to be all about user-centred knowledge system design, and Linked Data, however the slides are all available with the programme for the workshop.
Once again, my sincere thanks to the conference organisers for my opportunity to participate in ECDL2010. The conference proceedings are available from Springer, for those who want to follow up further, and presentation slides are gradually appearing on the conference website.