I'm not saying it doesn't have a learning curve. But it has much more advantages than making a custom equals() unnecessary.

Can we close this thread now? Can we agree that any article that mentions the words "Hibernate" and "identity" in its title should explain the Hibernate identity map (the persistence context)?

Yes, I think it's time to finish this thread. I would like to summarize, and you can correct anything that's wrong:

I was originally concerned that you felt there was some factual error or problem with assigned ids within the context of detached objects. That appears not to be the case. The criticism mentioned is that I went off and enspoused a technique that solves problems that only arise when working with detached objects. Meanwhile I completely failed to mention the technique of extended sessions which eliminate the need to detach objects. Fair enough. I should have framed the context of the technique at the beginning of the article.

You've been promoting extended sessions for some time and seem frustrated that people are still choosing to detach their objects instead. Perhaps you should provide more links and direct references to the relevant material instead of just saying "read the book". Actually I've found the online material on the technique somewhat sparce and hard to find. This is compounded by the evolving terminology: what was originally "long session" (reference docs prior to 3.2 and HiA) is now "extended session" and "session-per-conversation" (reference docs 3.2, JPwH), which you've refered to here as "extended persistence context". Both books provide a much better description than any online material I've found.

To clarify in case anyone's interested:
Sparce description in the Reference Documentation:
http://www.hibernate.org/hib_docs/v3/reference/en/html_single/#transactions-optimistic-longsession

Hibernate in Action: section 8.2.4

Java Persistence with Hibernate: section 11.2.3

Section in the wiki on the open session in view pattern:
http://www.hibernate.org/43.html#A5

Please extend this list with anything else you think people should read. I'll definitely add this summary to the article forums to make sure people have access to as much information as possible.

Cheers,
James

There are many more resources and more documentation that explain "long session". There is not a single Hibernate concept that is better documented. I don't have the time to make a list for you.

One is rather riding high this evening isn't one... don't fall off.

There is one last thing big drawback concerning application assigned identification.

What if your database is changed outside your Java program? What about distributed applications? Your database would become corrupt. That is why database assigned ID's are so much better then application assigned ID's.

And don't even get me started about having to debug a database full of UUID keys........ That just won't do.

It's my opinion that from a database design point of view using application asigned ID's as primary key's is poor design. Using UUID's is even worse.

I was missing this rather important point in the whole 'discussion' so I felt that I had to mention it.

I just read the article on the OReilly website and found it enlightening.

Even if its headline and conclusion are wrong, it still was well written and explained well some important fundamentals of coding against a persistence store.

Now, if someone has some patience left to bear with my two cents "this" puzzle.. that'd be sweet :-)

viewtopic.php?f=1&t=1011804