jackrabbit-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Nicolas Dufour" <nrduf...@gmail.com>
Subject Re: Question about node type changes
Date Mon, 13 Oct 2008 18:54:51 GMT
Just a piece of advice Alfie: DONT use nodetypes unless you absolutely need

I mean, if you are sure 100% that your content will never ever change, then
ok define a nodetype and its properties.
If you think your content will change a bit from time to time and you need
to classify it by nodetype then define a nodetype with nr:unstructured as

For any other cases: use nt:unstructured !


On Tue, Oct 7, 2008 at 4:05 PM, Brian Thompson <elephantium@gmail.com>wrote:

> In a way, Jackrabbit already supports a 'lax' mode; just use
> nt:unstructured
> for all your nodes.  No structure will be mandated.
> -Brian
> On Tue, Oct 7, 2008 at 2:19 PM, Alfie Kirkpatrick <
> Alfie.Kirkpatrick@ioko.com> wrote:
> > I'm new to Jackrabbit and have read with interest some of the archive
> posts
> > and JIRAs on the evolving support for changes to node type definitions
> when
> > there are active nodes using those definitions. These seem to assume as
> > fundamental that a node shall always conform to its node type.
> >
> > From a site development point of view where Jackrabbit might be the
> content
> > repository, this causes some concern to me. Sites change, requirements
> > change. To me the node type definition should be more of a guide to the
> > intent for the structure of the node rather than enforce very strict
> > validation. To me it's a bit like the question whether an XML editor with
> > DTD/schema validation should ever allow an invalid document to be
> created,
> > even temporarily. Most end up taking the approach that it's sometimes
> > necessary to make a document invalid on the way to making it valid again,
> > and this seems reasonable to me.
> >
> > So my question is really whether Jackrabbit is ever likely to support a
> > 'lax' mode where node types can be changed even if this causes existing
> > nodes to become invalid, or whether it's part of the fundamental design
> that
> > this should not be possible, ever.
> >
> > Apologies if this goes over old ground and thanks in advance for your
> > interest.
> > Alfie.
> >

Nicolas Dufour

  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message