Replacing betwixt, beanutils, and digester with castor

[Jason van Zyl <ja...@zenplex.com>]

Hi,

I would like to generate a blazingly fast SAX parser using Castor and
remove our dependencies on betwixt, beanutils and digester. I want to
generate the model from the XSD for now (explanation below) so this
would mean that validation would always work and the parsing would be
about as fast as it could get. Do some more research but we could
probably get away with using a minimal parser as you don't need a
validating parser if you generate castor sources with built in
descriptors.

It's something I would like to start as I know performance is of general
concern.


-- 
jvz.

Jason van Zyl
jason@zenplex.com
http://tambora.zenplex.org

In short, man creates for himself a new religion of a rational
and technical order to justify his work and to be justified in it.
  
  -- Jacques Ellul, The Technological Society

Re: Replacing betwixt, beanutils, and digester with castor

[Colin Sampaleanu <co...@exis.com>]

Jason van Zyl wrote:

>On Tue, 2003-02-25 at 08:54, Jason van Zyl wrote:
>
>What I meant to explain in the last message is that I don't ultimately
>want to base Maven's model in an XSD file. Ultimately I would like to
>use XMI directly. Currently I have stripped out the forward/reverse XMI
>tools from Argo and I'm going to put it all together in a plugin. I'm
>also chatting with the AxGen folks who make a tool that takes an XMI
>model and it lets you generate artifacts on the model using Velocity
>templates. So ultimately I want to generate Maven from XMI but will
>probably be a while. I will settle for the XSD for now.
>
>But for Tambora I'm taking an XSD, generating sources from Castor and
>then reverse engineering the XMI. Then with the XMI I can generate OJB
>repository metadata as that was one of the first goals of the AxGen
>project. But they have templates for DDL generation and OJB proxies and
>class generators. That's all I wanted to explain.
>
>  
>
Btw, another source of freely available source for XMI and code 
generation is the Eclipse Modelling Framework (EMF) project from Eclipse:
  http://www.eclipse.org/emf/
I have not looked at EMF specifically, but the Eclipse code is generally 
very well thought out and well tested.

Re: Replacing betwixt, beanutils, and digester with castor

[Jason van Zyl <ja...@zenplex.com>]

On Tue, 2003-02-25 at 08:54, Jason van Zyl wrote:

What I meant to explain in the last message is that I don't ultimately
want to base Maven's model in an XSD file. Ultimately I would like to
use XMI directly. Currently I have stripped out the forward/reverse XMI
tools from Argo and I'm going to put it all together in a plugin. I'm
also chatting with the AxGen folks who make a tool that takes an XMI
model and it lets you generate artifacts on the model using Velocity
templates. So ultimately I want to generate Maven from XMI but will
probably be a while. I will settle for the XSD for now.

But for Tambora I'm taking an XSD, generating sources from Castor and
then reverse engineering the XMI. Then with the XMI I can generate OJB
repository metadata as that was one of the first goals of the AxGen
project. But they have templates for DDL generation and OJB proxies and
class generators. That's all I wanted to explain.

-- 
jvz.

Jason van Zyl
jason@zenplex.com
http://tambora.zenplex.org

In short, man creates for himself a new religion of a rational
and technical order to justify his work and to be justified in it.
  
  -- Jacques Ellul, The Technological Society