jackrabbit-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Christophe Lombart <christophe.lomb...@gmail.com>
Subject Re: Advice on building a CMS
Date Mon, 30 Jan 2006 10:23:19 GMT
Hi Marteen,

On 1/29/06, Maarten Volders <maarten.volders@telenet.be> wrote:
> - Is it a good idea to create classes like Page, Paragraph, Content,
> Template, ... which hide direct calls to the node. So for example a call
> like page.getTitle() is actually a node.getProperty("cms:title"); I have
> not yet have a lot of experience using node operations so I do not know
> if I'm just working myself in trouble using such a approach or aren't
> there any issues with such a approach?

Yes, this kind of idea is quite new :-) but certainly this is an
interesting one.
In the Graffito project, we are working on this kind of mapping.
We are currently working on the documentation (and the site)  & the
advance mapping features. So, you will find more information on
http://incubator.apache.org/graffito/ in a couple of days.
Basic mappings work fine without important issues and now we are
analysing mapping with inheritance, interfaces,  complex compositions,

Of course, it is based on spring. You can see also in the Spring
module project, a nice JCR support.

> -I would like to build a web user interface that shows the tree
> structure of the pages in my cms. But I do not want that all the pages
> would be loaded immediately. But by using dwr / ajax I would like to
> have async calls to my service layer and retrieve the pages / nodes I
> need only when opening a page in the tree structure. Has anyone got
> experience with nodes in combination with dwr or ajax or other async
> calls. Issues I should now or some advice on best practices?

Building a treeview which can expand on demand is a nice approach but
there are plenty of ajax frameworks on the market and it is quite
difficult the best one. On  my current project, we are using Tacotine
without any problem. Thanks to Taconite, we are calling from the
browser some application services which are using our persitence
manager. This one returns a object graph matching to the JCR structure
to retrieve. There is a complete abstraction between the interface and
the JCR repository.

> - Also I was born in the age were there was no speak no more of
> hierarchy databases ... relational by all means they thought us. So I'm
> not very familiar on how the pages, paragraphs, content should be
> stored? Can I just have a root node wit under that page nodes, and each
> page node can have another page node under it or paragraph nodes. And
> paragraph nodes can have content and so on .... Workable solution or ... ?

Of course, The JCR model is very flexible. You can store any kind of
content. As usual, take the time to define correctly your model
(custome node types).

Jackrabbit gives the possiblity to store in a DB, file system or
another kind of persistence store.

> Other people wh have already build a cms or dms using jcr .... I would
> be great to get some input from you guys. Not only the technical stuff
> but also what experiences you have building user interfaces for jcr.

For simple cases (eg. readonly in a browser), you can use directly the
JCR API but when more business rules are imply, business objects
(pojo) are more appreciate. This is just my point of view and
certainly a lot of people on this list are not agree :-)

Best regards,


View raw message