Given the wide range of approaches for quantifying and propagating uncertain information, it would be a logical exercise to attempt to combine their capabilities into some kind of Grand Unified Theory of Uncertainty. Unfortunately, although there have been a number of attempts, such a unified representation does not currently exist and is unlikely to any time soon. A definition of a universal uncertainty interchange format would have to be able to guarantee compatibility with existing uncertainty models. In addition to possessing a superset of the representational expressiveness of each of the individual approaches it would also need mathematically rigorous methods for translating the qualities and quantities of uncertainty between its own representation and those of each of the separate approaches.
In order to deal with the lack of a unified theory of uncertainty, one of two primary approaches are typically employed.
• Accommodate only a single theory -has limitations in terms of dealing with different types of uncertainty and is especially difficult if different types of reasoning technology are involved
• Take an eclectic approach wherein different approaches are used in various parts of a system - Can have problems propagating the various forms of uncertainty between incompatible subsystems.
Although, as yet there is no single universal approach to handling uncertain information, that does not mean there is no point in dealing with the issue at all. Ignoring the problem does not make it go away. If handled correctly, even minimal accommodation for uncertain information can reduce the brittleness of a system or can provide an added vector of trust associated with the generated results. Over time techniques for dealing with uncertainty will continue to evolve and improve. In the meantime, even baby steps can provide value.