We've been using the OCM libraries with a certain amount of success.
We have heavily leveraged using the spring modules for configuration
and have used the repository pattern with a few base classes which
handle most of the OCM specific implementations. Our domain model maps
fairly cleanly into trees and nested structures, as I suspect many
models would, and we have done some amount of data normalization.
Adding JPA support to the OCM code would be a nice benefit for those
that have existing models, I'll grab the api docs and some sample code
and take a look at how big a change it would be. At first glance we
would have to provide a few implementations, but I think we have most
of the pieces they just need to be assembled. I have been working on
the entity factory with a second level cache, but got pulled away due
to "real" work stuff ;-)
In general, I feel that having two ocm style projects for such a small
group is detrimental, perhaps we could convince the jcrom folks to
come over to apache?
-paddy
On Feb 5, 2008, at 11:38 AM, Jukka Zitting wrote:
Hi,
On Feb 5, 2008 8:53 PM, Christophe Lombart <[EMAIL PROTECTED]
> wrote:
Concerning JPA, we can make a small study to check if JPA is really
a good
solution for a JCR backend ... if someone have time :-)
One very interesting use case for that would be to make Apache Roller
run on top of Jackrabbit. Roller is using JPA for persistence, and
AFAIK their content model doesn't contain anything that a hierarchical
content repository couldn't support.
BR,
Jukka Zitting