This section bridges from the challenges and opportunities described above, to the technology consequences and research questions that require answers.
Centralized and distributed query across persistent semi-structured metadata stores is of increasing importance in information systems. Semi-structured stores are growing in importance because the number of information-producing devices, human communities, and the digital assets that they produce, is exploding. Consequently, individuals and communities are not able to predict a-priori the number, purpose, or structure of individual structured schemas that will be required to support the creation of useful services and applications. Semi-structured stores are also increasingly important because the structure of community-based information is emergent and responsive to innovation. For services and applications to be successful they require the ability to use and draw relationships among elements in various schemas that were produced at different times, by different communities, with support for various application-enabling metadata. These schemas might be instantiated for a particular digital asset and populated by individuals contributing to a community, by individuals collaboratively maintaining a corpus on behalf of a community, by devices automatically, by supportive mining algorithms, or some combination of these. Innovation in any of these areas brings with it the need to evolve schemas. In this environment systems must be able to flexibly balance specialization and convergence. Specialization responds to the need to support emergent structure and schema evolution in response to innovation. Convergence responds to the need to stabilize schemas and provide pervasively understood mechanisms for discourse about assets.
The amount of information stored in such systems is set to grow exponentially for the foreseeable future. Many argue that the structure of information in such systems will be described by a large number of community-specific schemas. Users of such systems will arguably want and need to be able to store and manage information in a variety of interoperating semi-structured repositories tuned and tailored to individual, community, and/or institutional needs and requirements.
Note that a standard query language for persistent semi-structured stores does not yet exist. Neither exists widely accepted user interface designs and metaphors that will allow end-user humans to effectively interact with such systems.