Overview:
Precise specifications of semantics are essential to understand business
rules and to create software systems satisfying these rules. The needs
of subject matter experts have to be understood by the modelers and conveyed
to the developers; and the developers have to satisfy precisely those needs,
no more and no less. The ultimate goal is to solve business problems rather
than just to produce code.
In the same manner as software patterns are used in software systems architecture by abstracting code, business patterns are (re)used to elicit and specify business requirements by abstracting collections of business entities, together with their relationships and behavior. Isolated objects are not semantically rich enough for reusability; and early allocation of behavior to classes is often not understood by business customers. The concerns of the business enterprise should be clearly separated from the concerns of the system development; and correspondence between these should be explicitly specified by means of a linking invariant.
Declarative specifications of behavior using invariants and pre- and postconditions is well-known and may be consistently used at all stages of the information system life cycle. Although we seem to know how to write good specifications and (better!) how to write good code, we still know substantially less how to bridge the gap between the two.
In an open system environment, properties of objects, their relationships and behavior are not fixed. Requirements change. Even invariants may change with time. New viewpoints on existing constructs can be discovered. New properties may contradict the existing ones, and so there is a need to specify explicitly which property values have priority. Thus, non-monotonic reasoning becomes of direct business relevance rather than an esoteric academic exercise. Proper business specifications formulate the problem and thus are essential for discovering the solution.
Requirements for Attendance:
Submissions (4-6 pages) on these and related issues are due before
August 5, with notifications of acceptance by August 20. Electronic submissions
are strongly preferable. Accepted submissions will be distributed before
the workshop.
Organizers:
Haim Kilov, IBM T. J. Watson Research Center
Bill Harvey, Institute of Information Management, Robert Morris College
Submissions:
Haim Kilov
IBM T. J. Watson Research Center
30 Saw Mill River Road
Hawthorne, NY 10532
Email: kilov@watson.ibm.com
