We have worked with C++ for the past 5 years, but are new to EJB
design/implementation. One of our biggest challenges is knowing when to map C++ design
paradigms that we are familiar with (signeltons, polymorpism, ...etc) to EJBs and
knowing when we should be thinking in terms of different paradigms.
For instance, there are a number of places in our design where we have "manager"
objects that encapsulate a collection of objects from which they select an object to
delegate work to. Such "manager" objects are logically singeltons, however, EJBs seem
not to provide a direct representation for singeltons. This leads to a number of
related questions:
1) Is this design approach reasonable? Are there any EJB patterns that might apply?
2) We have seen several approaches to implementing singletons mentioned in this
newsgroup. The first is creating an object bound using JNDI, and the second is
creating an object with a static accessor. Which approach is generally preferred?
What are the pros and cons?
Does using a JNDI based object pose the risk of a single point of failure or scaling
problems?
3) When should the singleton objects be created and initialized? Should this always
be done on server startup?
Any and all input is greatly appreciated. Thanks!
Seth Hawthorne
OpenGrid
===========================================================================
To unsubscribe, send email to [EMAIL PROTECTED] and include in the body
of the message "signoff EJB-INTEREST". For general help, send email to
[EMAIL PROTECTED] and include in the body of the message "help".
