The 'wish' of every computer is to handle well defined 'crisp' data. Our everyday use of time contrasts these needs in an obvious way. Terms like era, age, epoch or season are most of the time ill defined. Sometimes crisp definitions were proposed, but they often have two drawbacks which prevent a wider application. Firstly, numerous people have tried to provide crisp definitions, resulting in a variety of statements. Secondly, a lot of temporal aspects simply cannot be defined crisply due to their nature. For example the four seasons summer, autumn, winter and spring are defined as three-month periods. They are often used as a temporal raster in the analysis of biological data. The intuitive definition of these seasons is somehow related to temperature, weather and vegetation conditions which are different in each season. But they are not constant over the years. The use of three month periods for the definition of seasons becomes even more difficult when traveling from north to south over the globe. The seasons are different in Europe than in Australia or in the Tropical Zone and have smooth transitions in between these areas.
This is only one example of temporal terms often used in our languages and in research. There are relative terms used to describe temporal relations such as before or after, lately or recently. The first aspect of these terms stating that it is in the past or future is well defined, whereas the second aspect that there is a close (temporal) proximity to the reference time is only vaguely expressed. This becomes even more pronounced in terms like while, during and contemporary.
In biology such terms are also widely used. Breeding season, migration time, generation time or descriptions of the age of an animal like juvenile or adult are only a few examples.
There are at least two ways to improve this situation for computational purposes. The first is to close the semantical gaps in the terminology of temporal aspects. Several authors (e.g., Korte, 1997) discussed various forms of temporal relationships between two objects. Nevertheless, a coherent framework and the corresponding semantics are still missing for efficient communication and application in computers. The second improvement might build on the ongoing research on fuzzy sets, boundaries and objects in the spatial domain (Burrough and Frank, 1996; Schmitt, 1996; Zadeh, 1965). Such concepts are also needed in the temporal domain as illustrated by the examples above but have not been applied yet.