“…Brute force approaches to these large collections (Cacheda, Plachouras, & Ounis, 2005) compete with methods to tease structure and genre out of scanned images of texts or full‐text documents, to increase precision in retrieval, to improve understanding of context, and to exploit linkages between elements of the structure of documents (Muehlberger, 2002; Rauber & Merkl, 2003). Distributed architectures and access using content‐based image retrieval (CBIR) (Tang, Avula, & Acton, 2004) are still attracting substantial research for massive image collections but these high‐tech solutions seem oddly less successful than allowing user annotation in collaborative workspaces (Pisciotta, Dooris, Frost, & Halm, 2005). Music, too, can be retrieved automatically with waveforms (Clausen, Kurth, Maller, & Ribbrock, 2004), but evidence suggests that organizations which take into account work processes might better support regular users (Notess, Riley, & Hemmasi, 2004).…”