Well, what is this likely to mean for digital libraries and digital library technologies? An oddly Janus-like scenario, perhaps. First, it argues for technologies that are modular, extensible, and easily customized for use in diverse and specialized settings. It also argues for ancillary tools that will help us achieve specialized uses – sort of an analogue to the machine tool technologies that remain a robust sector in the modern economy. Second, it suggests that tools that help users manage information from multiple sources at the desktop will become increasingly important. This will apply to many of the perennial themes in information/digital library research. For example, information retrieval and management tools that work for individuals and small groups of users, achieved by a convergence of research typically housed under "user interfaces," "information retrieval," "search," "filtering," "intelligent agents," and "groupware."
These two implications address the increasing granularity of digital library technologies – things that have to work at a relatively small or intimate level rather than at the scale of the Web, as we see it expanding. But there is also the power of the digital library technologies to achieve scale. This goes back to the original notion of resource sharing that drove early projects in networking and parallel processing. From a digital libraries perspective, the efficiency here obtains from the idea of store once/use often. That is, large and complex materials with perhaps limited or occasional use can be stored once but accessed often. This application of digital library technologies is closer to notions of infrastructure – items that are shared – and has associated with it difficult questions of threshold costs and responsibilities as well as some specific issues of remote searching, data partitioning, and transfer.
Many of us – myself included – are still in love with the vision of a networked world where everything "talks" to everything else. It may well turn out, however, that the parts will become at least as important as the whole.