axis-java-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Jarek Kucypera <...@mikrozet.wroc.pl>
Subject Re: Axis2 1.2 sanity check: Can generated POJO service/client handle 0-length array?
Date Tue, 12 Jun 2007 11:11:14 GMT
Anne Thomas Manes wrote:

> I trust you recognize
> the value of clean separation between your application object model
> and a database data model and the DAO design pattern. 

Of course I do. My point is that I wanna write the interface layer
of my system in Java, not int WSDL, of course for the price of not using
all of the wsdl/schema/xml features.
Actually we are doing it in our current project: we created
core engine for the business logic and a separate,thin interface layer.
The interface is simple enaugh to write it as a set of
of beans and service classes whose only task is to translate params/
results and call the core. Due to simplicity of the interface and the fact
we don't need any of  additional ws features like addressing,
reasonable POJO support just fits this scenario.

> There is an impedance mismatch between XML data structures and Java
> object models, just as there is an impedance mismatch between SQL data
> models and Java object models. In many cases the mismatch is
> insignificant, but it's a bad idea to assume that it doesn't exist.
>
Noone does assume that. And in those "many cases" it's worth to pay the 
price
of simplicity to avoid diving into wsdl.

> I agree that Axis2 could handle POJO deployment better, but even if it
> were the best that it could be, I seriously doubt that it could handle
> everything automatically.

I don't  expect POJO to handle everything,.
But reasonable propagation of exceptions is a minimum for it to be usable.

J.K.

---------------------------------------------------------------------
To unsubscribe, e-mail: axis-user-unsubscribe@ws.apache.org
For additional commands, e-mail: axis-user-help@ws.apache.org


Mime
View raw message