Submitted to a conference. Full paper (htm).
We report on a prototype system that provides access to executable predictive tools, as simulations and spreadsheets. We also discuss its interoperation with SQL accessed resources which will augment the decision-making support capabilities of information systems. The central component is a new interface language, SimQL, which mirrors the functionality of SQL, but delivers information projecting future states, obtained from a variety of simulations.
Simulations to be wrapped for SimQL access range from spreadsheets, via business and governmental planning models, to large remote continuous simulations, as used for weather forecasting. Results reported through SimQL are paired data elements, the expected value and its certainty. SimQL is intended to be used within information systems that cover data from the past into the future, and support the assessment of the effects of alternate decisions, so that multiple future courses can be compared. Placing results of simulations into a consistent framework with databases and web-based information, avoids the system inconsistencies that decision-makers face today.
The long-range motivating vision is that an interface language provides separation of customers and tool providers. Their autonomy will allow information consumers and providers to make progress independently, mirroring the past decades of SQL use.