Hello Simon,
Simon Pamies wrote:
Hallo Bodgan,
here at banality we're currently evaluating possibilities to create a
content management system based on Java technologies like Jackrabbit
because we want to switch away from Plone based sites (very long term
planning). We have built our own heavy-customized Plone but it doesn't
longer fit our needs.
Surfing the internet I discovered that nuxeo's ecm is built on top of
jackrabbit and I wanted to try out your implementation
(org.nuxeo.ecm.core.jcr.*). On the one hand we could build our own
wrapper around the JSR API but on the other hand it would be easier to
concentrate on other ideas we have and use existing wrappers. I saw
that your code is GPL based and so thought about reusing your stuff.
We will also release our system using an open-source license and
perhaps we can share ideas and implementation.
So here my questions:
a) Do I have to download the complete org.nuxeo.ecm.* hierarchy in
order to get the jcr stuff running?
No. You need only the Nuxeo Runtime and Nuxeo Core layers. This means
you need all nuxeo runtime projects
from http://svn.nuxeo.org/trac/nuxeo/browser/org.nuxeo.runtime/trunk
and also all nuxeo core projects from:
http://svn.nuxeo.org/trac/nuxeo/browser/org.nuxeo.ecm.core/trunk
For now you can run these on jboss - the supported jboss version is
specified in Developer Quick Start guide
http://svn.nuxeo.org/trac/nuxeo/wiki/DeveloperQuickStart
So you can install nuxeo5 as usual and then remove all jars from the
nuxeo.ear but the ones in system and core directories:
As a remark - nuxeo core is also working on glassfish the reference
application server (made by sun)
but there are not yet build scripts provided for this - we are working
on this.
b) Are there any documentation about your wrapper and the repository
structure you've defined?
We don't have a documentation yet that covers the JCR mapping - but you
can get help by posting questions on the nuxeo ecm list.
In the source code you may find some info about this.
So please register on the Nuxeo ECM list to get help ;-)
Roughly, we define a document model that is mapped over a subtree of JCR
nodes something like this:
+ Root (the document node)
(scalar properties are mapped as JCR properties directly under the
document node)
- scalar_property1
- scalar_property2
- ...
(complex properties are mapped as JCR nodes directly under the
document node)
+ complex_property1
+ complex_property1
+ ...
(the container of the children document nodes - each child document
will be mapped in a child node under this node)
+ ecm:children
We also use XSD to define Document types. Each of these schemas are
mapped to JCR node types (using a user defined namespace)
The document types and schemas, the document model, the query model,
searching, versioning and all these are covered by nuxeo core (nuxeo
core is the name of the nuxeo repository layer).
The nuxeo runtime is the framework on top of which nuxeo components
(like those forming nuxeo core or nuxeo platform) are running
As a remark, the nuxeo core is also working outside an application
server - for example in a desktop application as an embedded repository
This is covered by the Apogee project - but for now there we don't have
an Apogee that works with an embedded repository - this will come soon
Regards,
Bogdan
Thanks!
--
Simon Pamiés
CTO - Programmierung
Geschäftsführender Gesellschafter
banality - design & communication
Gadderbaumer Straße 3
DE-33602 Bielefeld (Germany)
+ 49 (0)5 21/7 84 88-27 (tel)
+ 49 (0)5 21/7 84 88-25 (fax)
[EMAIL PROTECTED]
www.banality.de
_______________________________________________
ECM mailing list
[email protected]
http://lists.nuxeo.com/mailman/listinfo/ecm
