Mailing-List: contact users-help@jackrabbit.apache.org; run by ezmlm
Precedence: bulk
Reply-To: users@jackrabbit.apache.org
Received-SPF: pass (athena.apache.org: domain of christophe.lombart@gmail.com
 designates 74.125.44.28 as permitted sender)
DomainKey-Signature: a=rsa-sha1; c=nofws;
        d=gmail.com; s=gamma;
        h=message-id:date:from:to:subject:in-reply-to:mime-version
         :content-type:references;
        b=q+JVCrudslu9GxgeJRNQYkIqJL9VDEt5MGo7+BDsz/fxeGElYYGVxDb6a1c51yYHzj
         8YvPbOQu/oPUFN86b2YDX3bz1tWP9tpc2hRs1P1MXEj4pwyv9FJP1w4O0dLV7KSp/Ytn
         TcSo0eH4R1/QfLTXONRXIWu86HHC/hkpcdIQI=
Message-ID: <3b728ee90811110649p7b8b935bwa5c37c6b6953903a@mail.gmail.com>
Date: Tue, 11 Nov 2008 15:49:52 +0100
From: "Christophe Lombart" <christophe.lombart@gmail.com>
To: users@jackrabbit.apache.org
Subject: Re: an alternative OCM (Object Content Mapping)
In-Reply-To: <200811111327.58369.lukin@stu.cn.ua>
MIME-Version: 1.0
Content-Type: multipart/alternative;
	boundary="----=_Part_99291_15213255.1226414992465"
References: <20414717.post@talk.nabble.com> <4919573C.9030804@gmail.com>
	 <E599B819E62E8C4F85F610411F0E38A077EEAF@PUNEXCH.bioimagene.com>
	 <200811111327.58369.lukin@stu.cn.ua>

------=_Part_99291_15213255.1226414992465
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit
Content-Disposition: inline

On Tue, Nov 11, 2008 at 12:27, Aleksei Lukin <lukin@stu.cn.ua> wrote:

> On Tuesday 11 November 2008 12:29:46 Boni Gopalan (BioImagene) wrote:
> > OCM supports annotations and XML mappings.  And it supports custom
> > converters for simple types, Beans and collections.  The only
> disagreement
> > I have with the design is the need to carry 'path' as part of the bean
> that
> > I want to persist.
> [skip]
> >
> > Thanks
> > boni
>
> Well, my 2c  about things I do not like in current OCM.
>

You are welcome.


>
> 1. ocm_discriminator node type registration at startup and other messy
> stuff.  I could be hidden in some hard-coded static function lile ocmInit().


Agree. From 1.5, this node type is not used. So, it is not necessary to
register it on startup.


>
> 2. path as primary key. I discussed it a lot with author and we are both
> wih our own opinion :) I use path in base class just to satisfy ocm, but for
> real key I use uuid. It is
> fast and it allows to keep reference consistance.


You know my position :-) but it should be nice to see the point of view from
others. For me Path is still a better canditate for an object ID. Maybe we
can try to support both (Path or UUID) as the object id.

>
> 3. class registration in annotation mapper. I think global mpper cold
> handle this and @Node annoation may put registration code into. I'm
> constatntly forgetting to add class
> there :)


Same problem here :-) . I'm still looking for a good way to scan one or more
packages to find annoted classes on startup. Do you know how I can do it ?


>
>
> Thing I miss is Map. But I guess it is under development now.


Map & collection supports are better . Some thing like this :

@Node Class MyClass
{
       @Collection List<AnotherClass> mylist;
       @Collection Map<String, YetAnotherClass> myMap;
}


This is really simpler compare to the version 1.4. Don't blame me if the
@Collection is not a good name. There is a jira issue to review that. The
JPA annotations OneToMany or OneToOne are better for me.


>
> May be it's time to think on some improvemnent of smart mapping.


Please create jira issues at least to give us some ideas. I would like to
implement a "default" mapping (without annotation and xml description).


>
>
> Rest of annotation based stuff I like, I use, I'm happy with and thank you,
> Christophe! :)
> --
> SY, Alex Lukin
> RIPE NIC HDL: LEXA1-RIPE
>

------=_Part_99291_15213255.1226414992465--