User:MadScientistX11/FramesVsObjects

Comparison of Frames and Objects edit

Frame languages have a significant overlap with object-oriented languages. The terminologies and goals of the two communities were different but as they moved from the academic world and labs to the commercial world developers tended to not care about philosophical issues and focused primarily on specific capabilities, taking the best from either camp regardless of where the idea began. What both paradigms have in common is a desire to reduce the distance between concepts in the real world and their implementation in software. As such both paradigms arrived at the idea of representing the primary software objects in taxonomies starting with very general types and progressing to more specific types.

The primary difference between the two paradigms was in the degree that encapsulation was considered a major requirement. For the object-oriented paradigm encapsulation was one of the if not the most critical requirement. The desire to reduce the potential interactions between software components and hence manage large complex systems was a key driver of object-oriented technology. For the Frame language camp this requirement was less critical than the desire to provide a vast array of possible tools to represent rules, constraints, and programming logic. In the object-oriented world everything is controlled by methods and the visibility of methods. So for example, accessing the data value of an object property must be done via an accessor method. This method controls things such as validating the data type and constraints on the value being retrieved or set on the property. In Frame languages these same types of constraints could be handled in multiple ways. Triggers could be defined to fire before or after a value was set or retrieved. Rules could be defined that managed the same types of constraints. The slots themselves could be augmented with additional information (called "Facets" in some languages) again with the same type of constraint information.

Driven by the demands of developers, most of the later Frame languages (e.g. Loom, KEE) also included the ability to define messages on Frames.[1]

The other main differeniator between Frame and OO languages was multiple inheritance (allowing a Frame or class to have two or more superclasses). For Frame languages multiple inheritance was a requirement. This follows from the desire to model the world the way humans do, human conceptualizations of the world seldom fall into rigidly defined non-overlapping taxonomies. For many OO languages, especially in the later years of OO, single inheritance was either strongly desired or required. Multiple inheritance was seen as a possible step in the analysis phase to model a domain but something that should be eliminated in the design and implementation phases in the name of maintaining encapsulation and modularity.

References edit

  1. ^ Macgregor, Robert (August 13, 1999). "Retrospective on Loom". isi.edu. Information Sciences Institute. Retrieved 10 December 2013.