Mailing-List: contact cocoon-dev-help@xml.apache.org; run by ezmlm
Precedence: bulk
Reply-To: cocoon-dev@xml.apache.org
From: "Carsten Ziegeler" <cziegeler@s-und-n.de>
To: <cocoon-dev@xml.apache.org>
Subject: RE: [RT] Access to the object model
Date: Thu, 22 May 2003 08:37:52 +0200
Message-ID: <LKENJPKDAIKAPKHEOCMMGENIDKAA.cziegeler@s-und-n.de>
MIME-Version: 1.0
In-Reply-To: <bagarh$fi$1@main.gmane.org>
Importance: Normal
Content-Transfer-Encoding: 7bit
Content-Type: text/plain;
	charset="us-ascii"

Stefano Mazzocchi wrote:
> 
> I'm big time with Sylvain here.
> 
> IoC is also defensive programming: I don't give you access to stuff that
> you can abuse.
> 
> not *you*, Carsten, but others will.
> 
Agreed.

> In a future where cocoon development will be distributed in various
> places without knowledged people like us stopping all abuses when they
> start, we must distill this knowledge into the API or Cocoon will become
> an architectural mess (it already is, in some parts, in fact we'll have
> to cleanup the object model one day)
> 
Ok.

> >>So, what about implementing this Context thing then?
> >>
> > 
> > 
> > +1 from me ;-)
> 
> I'm sorry, but I'm still -1.
> 
> An avalon component should *NOT* have access to the cocoon object model.
> There should be something else in between glueing the two.
> 
> Now, prove me wrong.
> 
:) Sorry, I'm a little bit clueless about how this glueing could look like.
What do you not like about the following approach?
If a component implements Contextualizable it gets the Context object.
Via the context object it can get the object model by a defined key.

Or we could restrict the access to destinct objects in the object model,
like request and response.

What do you think?

Carsten