I'm currently developing a portal-style web application to tie together a number of independantly developed systems, hopefully to improve their accessibility to users and to boost uptake.

Each of these systems returns data in a slightly different format, and I've developed a framework to allow "translation" modules to plug in and interface with each of the external systems, returning data to a core application which is responsible for data presentation and user validation, among other things.

I've been experimenting with XML::Writer as a way of standardising the returns from each of the translator modules, hopefully to allow me to develop a standard interface that the core application can make use of. This module provides much of what I require, and coupled with XML::Parser I can quickly decode the information when it's needed.

This started to get me thinking about Java, and its implementation of beans, and my thinking that the Perl I've developed is essentially simulating that. I did some more searching on CPAN and found PerlBean, but I'm not sure whether using that would have any great benefits. In addition, I'm very much in favour of keeping an XML interface around, as some of the people who'll be working on the system in the future don't use Perl and a standard, non-proprietry interface would be of use.

I'd be interested to hear peoples' thoughts on this, especially concerning the PerlBeans module. Does anyone have experience with this, and, if so, how beneficial was it over simply simulating beans with XML? Any other thoughts and comments are welcome...