David Nuescheler wrote:
The only thing in JCR that is a little unusual when mapping to ORM...
well...
ORM meaning Object Repository Mapping, right ;)?
h�h� yeah that works too :)
... is versioning. Do we need to version all the previous object states ? Or
were you thinking about a mapping that doesn't use versioning ?
well, i think there are many application that that would requiry explicit
or implicit versioning of the data. in my mind the reason why versioning
is not taken into account in classical orm solutions is really because
both rdbmses and pojos do not directly address this issue.
i would personally would like to see this as possibily "an aspect" of an
object, to add the explicit versioning capabilities. so to me it would be
an extension to current orm solutions...
It is interesting, but I guess the data set can get pretty large :)
All the rest seems like a good fit though : locking, search,
referencing, etc..
right, and inheritance (with mixins possibly "aspects"?) ;)
since i have absolutely no experience with orm i am probably
way out here ;)
hehe yeah and let's just add some RDF on top of all this so that we can
retrieve POJO based on ontology processing :)
cheers,
Serge...
