Imre wrote:
> Consider a financial application where stocks (securities) are modeled
> with
> entity beans. Aggregate info about them (current value, combined
> performance
> or the set of portfolios they belong to) is still part of the domain (i.e.
> independent of the client looking at it). If you modeled this latter info
> through session beans and you had 1000 clients, you would replicate the
> same
> info in 1000 session beans whereas they could be shared just as easily.
>
[Chip Wilson]
The collections of lightweight summary objects we are discussing are
serialized and sent to the client requesting them, they do not remain on the
server as any kind of state. They contain the information that the actor
uses to identify a particular entity that they would like to work with,
typically. They also contain the PK of the entity bean that they represent.
These objects are not domain objects, but are designed to contain the
information needed by a particular actor in a particular use-case, and
therefore belong in the application model layer. If there are collections
of domain objects that need to be provided to the client (as there most
certainly are, including the examples you gave), then these collections
should be made available through the interface of the relevant entity bean.
===========================================================================
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".
