I recommend custom conversion whether with static factory methods or something more brute force, although it might depend a bit on whether you're round-tripping the xml. If you're not, and are going to be hitting attributes in the nodes frequently, it will probably pay to convert the node from xml to class instance, especially if you don't need all of the attributes. Some webservices return a lot of data per-item that I'm just going to ignore and a generic XML walker iterates all the nodes and attributes while a custom one can pick and choose the few things it might need.
________________________________ From: flexcoders@yahoogroups.com [mailto:[EMAIL PROTECTED] On Behalf Of Richard Rodseth Sent: Friday, August 15, 2008 5:07 PM To: flexcoders@yahoogroups.com Subject: Re: [flexcoders] XML Decoding Not yet, though I saw a reference to that in my Googling. I think I'm just going to hold my nose gently and make static factory methods that take an XML argument. I've been experimenting with the XML-backed idea and that's not too bad either. Harder to switch between XML and AMF I guess, but KISS applies too. Unless there's a good library out there yet to be revealed. On Fri, Aug 15, 2008 at 5:03 PM, Tracy Spratt <[EMAIL PROTECTED] <mailto:[EMAIL PROTECTED]> > wrote: Have you looked at the Flex 3 functionality that reads the webservice wsdl and generates code for strongly typed access? It was way too complicated for my needs but might help answer your questions. Tracy ________________________________ From: flexcoders@yahoogroups.com <mailto:flexcoders@yahoogroups.com> [mailto:flexcoders@yahoogroups.com <mailto:flexcoders@yahoogroups.com> ] On Behalf Of Richard Rodseth Sent: Friday, August 15, 2008 5:12 PM To: flexcoders@yahoogroups.com <mailto:flexcoders@yahoogroups.com> Subject: [flexcoders] XML Decoding What's the latest and greatest on XML Decoding? Reading the blogs and docs, it looks as though SimpleXMLDecoder + ObjectTranslator is one approach, but seems to rely on (deprecated?) XMLDocument and is not schema-aware. I've also considered defining some interfaces and providing XML-backed implementations that would point back to the nodes in the original DOM, perhaps with binding support and caching of the typed objects. But I think I'd just prefer a nice clean way to convert the whole structure to typed objects, and then throw the XML away. Any recommendations?