Mailing-List: contact dev-help@avalon.apache.org; run by ezmlm
Precedence: bulk
Reply-To: "Avalon Developers List" <dev@avalon.apache.org>
Message-ID: <3F9997C4.2020503@apache.org>
Date: Fri, 24 Oct 2003 17:21:08 -0400
From: Berin Loritsch <bloritsch@apache.org>
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US;
 rv:1.4) Gecko/20030624
MIME-Version: 1.0
To: Avalon Developers List <dev@avalon.apache.org>
Subject: Documenting the TreeBuilder stuff
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit

I started putting in some information that would at least document the
interfaces alot better based on what I saw on the list so far.  I put


Each type of ProcessingNodeBuilder or ProcessingNode interface is nicely
separated, and they do one thing and one thing well.  I really don't have
any issues with that at all.

The TreeBuilder interface is much less straightforward.  It mixes the
concerns of the meta information for the language type captured in the
config file with a bunch of processing things.  I think we should look
at separating the interface into more logical units.  That would help
make it more understandable.

For instance, we have a TreeBuilder which represents the language.  That is
nice and good.  However, IMO it has one responsibility: to create
ProcessingNodeBuilders for a particular file.  The root ProcessingNodeBuilder in
turn should have the responsibility of creating the ProcessingTree.
The ProcessingTree in turn is used directly to process requests.

So, in essence, we have a couple of roles here--which might even simplify
the implementation a bit:

* TreeBuilder: Create a ProcessingNodeBuilder tree.  (what goes on in the
                implementation should not be exposed in the interface)
* RootProcessingNodeBuilder: Create a Processor (which is implemented by
                              the ProcessingNodes)
* Processor:  The actual runtime system we use (the assembled sitemap so
               to speak).

Currently, it looks like there are implementation details exposed in the
TreeBuilder, mixed with runtime details exposed for the processing node
builders.

I dunno, though.  Do we want the TreeBuilder to only expose an interface
to get a particular Processor based on a URI???  I'm undecided.

-- 

"They that give up essential liberty to obtain a little temporary safety
  deserve neither liberty nor safety."
                 - Benjamin Franklin


---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@avalon.apache.org
For additional commands, e-mail: dev-help@avalon.apache.org