Would be really grateful for any help -

1) Now I am doing evaluation of several OMR products and Hibernate seems to be the best. But our project badly needs some caching. As far as I understand Hibernate supports 3 caching strategies for process (JVM) level cache - read-write, nonstrict-read-write and read-only. I understand the purpose of each strategy but could not find any description of implementation. For example, at what point records get added to chache, get locked etc., what happens when we use read-only strategy but try to modify an object....? What is implementation difference between read-write and nonstrict-read-write strategies....?

Whould be really nice if somebody can give me a link with good description.

Also can anybody please advise how to deal with distributed (cluster-level) cache synchronization problem?

2) I have the same problem as described in

http://forum.hibernate.org/viewtopic.ph ... criminator

can anybody please advise any workaround?

3) Does anybody know why Hibernate team insists on not using discriminator for loading collections from the table which contains records of dirrerent types? For some reason they offer to use "where" clause in collection mapping. It helps in the case of one-to-many collection but it doesn't help in the case of many-to-many (see 2 above).

The reason is

(a) performance
(b) mapping a single relational association to multiple object associations is evil non-use of polymorphism (just bad OO design), IMO

This is the quote from Hibernate manual (4.1.13):

"Finally, polymorphic persistence requires the declaration of each subclass of the root persistent class. For the (recommended) table-per-class-hierarchy mapping strategy, the <subclass> declaration is used."

But you said that it is a bad design... Probably I missed something...

I our particular project we have department class which aggregates collections of PermamentEmployee and Contractor classes. Those two classes have the common ancestor Employee and stored in the same table because they have very subtle differences...

In any case appreciate you help and sorry for being so persistent:))

I thought we were talking about collections??

Yes, I was talking about collections.

But I have elements of my collections stored in the same table (as long as they have common superclass). As far as I understand that is recommended approach. For example, I have abstract class Employee and instances of its subclasses PermanentEmployee and TemporaryEmployee are stored in the same table.

And I have a class Department which aggregates two collections - permanent and temporary employees.

Moreover, Department and Employee relationship is many-to-many.

In this case I have problem when I try to load Department graph