Jim C., > A private method cannot be extended, so does that mean it is final also?
<AFAIK> In java, yes. </AFAIK> Object-oriented theory includes the concept of a constant. A constant is an entity (e.g. field, method, or object) whose value is the same throughout its entire extent. (extent is time of existence). Nearly all object-oriented languages implement constants by enforcing an assign-exactly-once policy. java uses the 'final' modifier to denote the need for this enforcement. In theory, a private method, seen by all instances of that class, could be reassigned by other methods within the same instance, and thus need not be final. <AFAIK> In java, there is no way to create or modify a private method at run-time (i.e. within the extent of an object). Even private methods received from serialized objects had to be compiled from source code prior to run-time. java.lang.reflect.* allows read access, and some limited address access but not creation nor modification. </AFAIK> <AFAIK> Consequently, since a java private method must be assigned its value at compile-time and cannot be modified at run-time, a private method must have the same value throughout its extent, and therefore be final. </AFAIK> In other languages (e.g. perl), that permit run-time reassignment of private methods, such self-modifying code is a security and maintainability issue. In object-oriented theory, constants represent a dependency (e.g. standards compliance, physical measurements, ...) outside the scope of the algorithm expressed. The use of constants is a trade-off between limiting the techniques available to the optimizer, and promising the optimizer that synchronization code can be safely removed. Attempts to seperate these two uses (and subsequently optimize both) have historically proven unsuccessful. <IMHO> When using constants to express dependencies outside the algorithm, the final modifier should be applied at the lowest level (field rather than method or class) possible, and the comments should point to the outside dependency (e.g. standard being complied with, source of physical measurement, ...). </IMHO> Hopefully helpful, -- Dr. Robert J. Meier ---------------------------------------------------------------------- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
