XML retrieval is a very active branch of Information Retrieval addressing the focused retrieval of semi-structured data. One of the main open problems in XML retrieval is to understand the information seeking behavior of real element retrieval users. Such understanding is crucial for further progress, since retrieval models, and even metrics, make explicit assumptions on user preferences---assumptions that are unwarranted by our current knowledge. In archival science, modern archival finding aids are encoded in XML using the Encoded Archival Description standard. However, up to now, archivists have focused almost entirely on the "supply" side rather than the "client" side: Web sites with digital finding aids often have no search functionality, and users often fail to locate the relevant archival material, frustrated by the intricacies of the lengthy and complexly structured finding aids.
The overall aim of the README project is
The proposed project will make dual contributions to both XML retrieval and to archival access. The user studies (key objective 1) will provide insight into user information seeking behavior in XML element retrieval, crucial for validating retrieval models and evaluation metrics. At the same time, these studies will provide crucial feedback on the strengths and weaknesses of digital finding aids for users in search of archival material. The developed retrieval techniques (key objective 2) will significantly extend current XML retrieval models by building on the results of the user studies, and by taking into account the user's profile and context, and the textual context of the unit to return. At the same time, the resulting search engine will greatly enhance the user access to archival material through digital finding aids.
A detailed project description is here.