Open World Design accepts that information exists in a global context that is evolving in unpredictable ways and avoids assuming that interoperability with future, possibly unanticipated sources can be more than partial. Open World Design stands in contrast to the Closed World Design of traditional IT environments, in which data is carefully controlled, with document formats and database schemas optimized for specific applications. Loosely specified vocabularies such as Dublin Core, when used with generic underlying models such as RDF, provide a good technological basis for Open World Design.
Since the early 1990s, the Web has placed local sources of information into a new global context, making it dramatically easier to pull information from multiple sources and giving rise to what is sometimes called an "open world mindset" for knowledge management.
The open-world mindset tolerates, even assumes, that the information at hand is incomplete and may be supplemented by incorporating information from other, possibly unanticipated, sources. Data integration is based on conformance not to fixed, ad-hoc data structures but to a generic model for expressing data. Data integration may be partial, based on descriptive attributes that may only partially overlap. New information can be incorporated without invalidating information already present. Systems can be rolled out early, with simple designs and focused on key, high-return requirements, then flexibly evolve as new sources of information become available and new requirements emerge.
Note that in this context, "open-world" refers to principles of application design. It specifically does not imply that the data is openly accessible ("open data"). Open-world design can apply equally well to proprietary data that is maintained behind corporate or organizational firewalls.