Hi,
Michael Wechner wrote:
David Nuescheler wrote:
i seriously think that the client app (content browser) should
be independent of the transport layer and just use jcr to
interact with the repository.
you mean by using jcr-rmi?
I think David suggested that the browser application should use JNDI or
some other mechanism to get a reference to the JCR Repository interface
and use the standard JCR methods to access the repository.
This way the application could easily be deployed using any of the
available deployment models. It could either directly access an
in-process Jackrabbit or another JCR implementation, or use JCR-RMI or
the Webdav layer to access a remote repository.
The choise of a transport layer and even whether to use one should be a
deployment, not development issue.
BR,
Jukka Zitting
