Interest in using UML for XML modeling tends to be cyclic. At times it has fashionnable, at other times it is. Some groups make it mandatory (e.g. ebXML), others just refuse to even touch it. This session proposes to shed some light on the subject by reviewing the pros and cons of UML modeling for XML, including criterias to identify those projects where it makes sense. I will draw from project experience to illustrate this. But more importantly, the session will discuss the tools. I believe that UML only makes sense in an MDA or MDA-like approach were the model is not just a pretty drawing but is the repository of information from which models, WSDL and other implementation aspects are derived automatically. I have been using UML since version 0.91 and I have been deriving DTDs (no schemas then) automatically from models since 1999. I have also participated in many projects that did not use UML modeling so I believe I have a fair, balanced view on the topic and enough project experience to draw samples from. The session will also draw from this series of article: http://www-128.ibm.com/developerworks/xml/library/x-wxxm23/